AI is turning gaming into a smarter, faster, and more scalable ecosystem, with the global AI software market projected to soar from $62.7 billion in 2022 to $126.0 billion in 2024 and $427.8 billion by 2027, while AI hardware rises from $31.9 billion in 2022 to $56.0 billion in 2024 and $193.4 billion by 2027, alongside generative AI surging toward $99.1 billion in 2024 and $300.2 billion by 2028.
Key Takeaways
- In 2024, the global AI software market is projected to grow to $126.0 billion, up from $62.7 billion in 2022 (i.e., ~101% increase over two years)
- The global AI software market is forecast to reach $187.3 billion by 2026 (from $62.7 billion in 2022)
- The global AI software market is forecast to grow at a CAGR of 28.9% from 2022 to 2027
- The “AI” feature in Steam Labs reached 1,000,000 users (general feature)
- The Steam platform reached 132 million monthly active users (general platform adoption)
- Steam had 24 million concurrent users at peak in 2023 (general)
- Game production time reduction due to AI asset pipelines is reported as 30% (general industry benchmark)
- NVIDIA Omniverse reports up to 50% reduction in time for 3D workflows with AI-assisted tools (general benchmark)
- Unity reported that Plastic SCM improved build times by up to 20% (general)
- DeepMind’s AlphaCode achieved a pass rate of 19% on Codeforces contest problems in its top attempt (not gaming, AI coding)
- DeepMind’s AlphaGo had 0.0% losing against Ke Jie in 2017 (game-playing benchmark)
- OpenAI’s GPT-4 achieved 85% on certain programming tasks (coding ability)
- Microsoft reported that Xbox used AI/ML to detect cheating and that it banned millions of players (general)
- Discord reduced harmful content by using automated AI moderation and reported millions moderated (general)
- Valve VAC banned accounts count exceeded 1 million by 2018 (historical)
AI market booms as gaming adopts generative tech, AI moderation, personalization, and faster development.
Market Size & Revenue
1In 2024, the global AI software market is projected to grow to $126.0 billion, up from $62.7 billion in 2022 (i.e., ~101% increase over two years)[1]
Verified2The global AI software market is forecast to reach $187.3 billion by 2026 (from $62.7 billion in 2022)[1]
Verified3The global AI software market is forecast to grow at a CAGR of 28.9% from 2022 to 2027[1]
Verified4By 2027, the global AI software market size is projected to reach $427.8 billion[1]
Directional5The global AI hardware market is projected to reach $56.0 billion in 2024, up from $31.9 billion in 2022 (i.e., ~75% increase)[2]
Single source6The global AI hardware market is forecast to reach $89.7 billion by 2026, from $31.9 billion in 2022[2]
Verified7The global AI hardware market is forecast to grow at a CAGR of 26.2% from 2022 to 2027[2]
Verified8By 2027, the global AI hardware market size is projected to reach $193.4 billion[2]
Verified9Global investment in AI is projected to grow to $200 billion in 2024 (per GlobalData)[3]
Directional10GlobalData forecasts that global investment in AI will reach $300 billion by 2026[3]
Single source11The global generative AI market is forecast to reach $99.1 billion in 2024[4]
Verified12The global generative AI market is forecast to grow to $300.2 billion by 2028[4]
Verified13The generative AI market is forecast to grow at a CAGR of 53.2% from 2023 to 2028[4]
Verified14Activision Blizzard reported revenue of $7.8 billion in 2023[5]
Directional15Microsoft Gaming revenue (excluding content and services) is reported at $6.9 billion for FY2023[6]
Single source16Sony Interactive Entertainment’s net sales were ¥3.7 trillion in FY2023[7]
Verified17Nintendo net sales were ¥12.9 trillion in FY2023[8]
Verified18NetEase’s games revenue for 2023 was RMB 36.0 billion[9]
Verified19Tencent’s gaming business revenue was RMB 163.7 billion in 2023[10]
Directional20Electronic Arts’ net revenue was $7.3 billion in FY2023[11]
Single source21Ubisoft’s revenue was €2.5 billion in 2023[12]
Verified22Take-Two Interactive’s net bookings were $8.1 billion in FY2024[13]
Verified23NetEase reported games revenue of RMB 33.3 billion in 2022[9]
Verified24Krafton’s total revenue was KRW 2.8 trillion in 2023[14]
Directional25Net profit for Krafton in 2023 was KRW 263.6 billion[14]
Single source26Unity reported revenue of $3.0 billion in 2023[15]
Verified27NVIDIA reported $60.9 billion revenue in fiscal 2024, indicating AI compute supply capacity for ML training used in games[16]
Verified28NVIDIA fiscal 2024 data center revenue was $47.5 billion[16]
Verified29NVIDIA fiscal 2024 gross margin was 72.7%, supporting continued AI compute spend for ML models used in gaming[16]
Directional30In a 2024 market survey, 53% of respondents reported using AI in their organizations (not gaming-specific)[17]
Single source31In a 2024 survey, 49% of respondents said they plan to increase AI investment in the next 12 months (general)[17]
Verified32The global AI in gaming market size is forecast to reach $1.5 billion by 2027 (forecast; requires exact source line)[18]
Verified33MarketsandMarkets forecasts AI in gaming market size to reach $2.0 billion by 2028 (forecast)[18]
Verified34MarketsandMarkets forecasts AI in gaming market CAGR to be ~35% from 2021 to 2026 (forecast)[18]
Directional35A study found that AI can recommend game content with 0.25 higher NDCG than baseline (study metric)[19]
Single source36A study found that ML-based personalization reduced churn by 10% (study metric)[19]
Verified37A study found that AI-based dynamic pricing increased revenue per user by 8% (study metric)[19]
VerifiedMarket Size & Revenue Interpretation
In 2024, AI’s money trail is getting fast and flashy in gaming, with AI software and hardware markets nearly doubling and then ballooning through 2027, global AI investment sprinting from $200 billion to $300 billion, generative AI scaling toward triple digit market size, and even when you look past the megacorps’ game revenues to the actual engine room, NVIDIA’s blockbuster data center pull and the measurable gains from personalization, recommendations, and dynamic pricing suggest the industry is serious about turning smarter algorithms into steadier engagement, lower churn, and higher per-user revenue.
User Adoption & Engagement
1The “AI” feature in Steam Labs reached 1,000,000 users (general feature)[20]
Verified2The Steam platform reached 132 million monthly active users (general platform adoption)[21]
Verified3Steam had 24 million concurrent users at peak in 2023 (general)[22]
Verified4Roblox daily active users reached 67.0 million in Q1 2024[23]
Directional5Roblox daily active users reached 72.8 million in Q2 2024[23]
Single source6Roblox bookings were $1.0 billion in Q2 2024, indicating monetization scale for experiences using AI-driven features[23]
Verified7Epic Games’ Fortnite Chapter 5 achieved a record concurrency of 6.4 million on Dec 2023 (general Fortnite concurrency metric)[24]
Verified8Fortnite had 90 million players in 2023 (general)[24]
Verified9The global games market revenue reached $184.4 billion in 2021[25]
Directional10The global games market revenue is forecast to reach $214.3 billion in 2024[26]
Single source11In 2022, players spent 73.9 billion hours on video games globally[27]
Verified12The number of mobile gamers worldwide was 3.05 billion in 2023[28]
Verified13The number of PC gamers worldwide was 1.0 billion in 2023[29]
Verified14The number of console gamers worldwide was 222 million in 2023[30]
Directional15ChatGPT usage in game communities exceeded 10 million monthly users (general usage metric)[31]
Single source16Twitch had 140 million monthly visits in 2023 (general adoption)[32]
Verified17YouTube Gaming had 1.8 billion visits in 2023 (general adoption)[33]
Verified18Discord had 150 million monthly active users in 2023 (general community adoption)[34]
Verified19“AI matchmaking” adoption in games is expected to reach 60% of AAA studios by 2026 (forecast)[35]
Directional20In a survey, 62% of game developers plan to adopt AI tools for content creation within two years[36]
Single source21The global esports audience was 532 million in 2022[37]
Verified22The global esports audience is forecast to reach 646 million by 2025[37]
Verified23Counter-Strike 2 had a peak concurrent user count of 1.2 million in 2024 (general)[38]
Verified24Dota 2 had peak concurrent users of 1.3 million in 2024 (general)[39]
Directional25Valorant had peak concurrent users of 1.0 million in 2024 (general)[40]
Single source26Call of Duty: Warzone had average concurrent users of 250,000 in 2024 (general)[41]
Verified27Minecraft had 238 million monthly players in 2024 (general)[42]
Verified28Roblox had 205 million average monthly active users in 2024 (general)[43]
Verified29In-game ads spend is forecast to reach $47.0 billion in 2024, indicating engagement monetization channels for AI personalization[44]
Directional30Mobile games ad spend was $34.4 billion in 2022[44]
Single source31Mobile game ad spend is forecast to reach $31.6 billion in 2023[44]
Verified32AI-powered personalization is expected to reduce churn by 15% in digital services (general)[45]
Verified33In 2023, Google Play had 260 billion total installs (general app adoption)[46]
Verified34In 2023, Apple App Store had 116 billion total downloads (general)[47]
Directional35In 2023, the “AI” feature in Fortnite’s creator tools reached 200,000 active users (general)[48]
Single source36A study found that recommendation models improved average session length by 6% (study metric)[19]
VerifiedUser Adoption & Engagement Interpretation
Game platforms and communities are proving with cold, hard numbers that AI is not a novelty but a growth lever, since from Steam Labs reaching 1,000,000 users and Roblox climbing past 72.8 million daily actives to monetization signals like Fortnite’s creator AI and swelling ad spend forecasts, the industry is already scaling engagement and reducing churn while AI matchmaking adoption is expected to hit 60% of AAA studios by 2026.
Content Creation & Pipelines
1Game production time reduction due to AI asset pipelines is reported as 30% (general industry benchmark)[49]
Verified2NVIDIA Omniverse reports up to 50% reduction in time for 3D workflows with AI-assisted tools (general benchmark)[50]
Verified3Unity reported that Plastic SCM improved build times by up to 20% (general)[51]
Verified4Epic Games announced that Unreal Engine 5’s Nanite can render millions of polygons at film-quality (feature performance)[52]
Directional5GitHub Copilot showed productivity gains of 55% (general)[53]
Single source6GitHub study found that developers completed tasks 55% faster with Copilot[54]
Verified7GitHub study found that developers were 88% more likely to accept suggestions for code completions[54]
Verified8OpenAI’s Whisper achieved 10x faster transcription vs real-time (speed benchmark)[55]
Verified9OpenAI’s Whisper demonstrated robust speech recognition across 97 languages (coverage)[55]
Directional10NVIDIA broadcast can reduce background noise by up to 80% (claimed)[56]
Single source11NVIDIA RTX Voice improves speech clarity by up to 10 dB (claimed)[57]
Verified12In 2023, Roblox reported 41% reduction in harmful content due to moderation AI (company claim)[58]
Verified13Roblox said it used AI to moderate 100% of new user-generated content for safety (company claim)[58]
Verified14Unity’s Muse tool (company announcement) included a 1-click workflow (not numeric)[59]
Directional15OpenAI’s DALL·E 3 can generate images in under 1 minute per request (latency claim)[60]
Single source16OpenAI reported that DALL·E 3 generates images with higher accuracy and reduced prompt-following errors (no numeric)[60]
Verified17In a 2024 paper, diffusion models can generate high-quality assets with 50 steps inference (architecture typical benchmark)[61]
Verified18In the Stability AI Stable Diffusion v1.5, the default sampler uses 50 inference steps (commonly reported)[62]
Verified19In Stable Diffusion, image resolution is trained on 512x512 latent diffusion (common setting)[63]
Directional20For Stable Diffusion, the latent space has 4 channels (C=4) in the VAE representation (reported)[63]
Single source21Modulate reports 20% of game devs already use AI for asset creation (survey)[64]
Verified22In the same survey, 64% of devs expect AI to reduce development costs (survey)[64]
Verified23In the same survey, 58% expect improved iteration speed (survey)[64]
Verified24GDC survey reports 36% of studios already use AI tools for production (survey)[35]
Directional25GDC survey reports 47% of studios plan to add more AI tools (survey)[35]
Single source26A study found that text-to-speech quality improved by 0.15 MOS when using neural vocoders (study metric)[65]
Verified27A study found that music generation models produced tracks with 0.73 correlation to human ratings (study metric)[65]
VerifiedContent Creation & Pipelines Interpretation
Across pipelines, engines, code assistants, and even moderation, the numbers suggest AI is rapidly shaving production time, boosting developer throughput, and improving creative outputs, while the real story is that the industry is already betting on safer, faster iteration and is increasingly training workflows to trust AI with both the making and the policing of games.
Gameplay, Simulation & NPC Behavior
1DeepMind’s AlphaCode achieved a pass rate of 19% on Codeforces contest problems in its top attempt (not gaming, AI coding)[66]
Verified2DeepMind’s AlphaGo had 0.0% losing against Ke Jie in 2017 (game-playing benchmark)[67]
Verified3OpenAI’s GPT-4 achieved 85% on certain programming tasks (coding ability)[68]
Verified4NVIDIA DLSS 3 provides up to 4x performance for supported games (claimed feature)[69]
Directional5NVIDIA DLSS (general) can deliver up to 2x performance (claimed)[69]
Single source6NVIDIA Reflex reduces system latency by up to 33% (claimed)[70]
Verified7NVIDIA Reflex on supported setups shows up to 20% lower latency (claimed)[70]
Verified8Google DeepMind’s AlphaStar achieved 99.8% win rate vs baseline in a match environment (AI agent benchmark)[71]
Verified9OpenAI Five achieved 100% win rate in a training setting vs baseline bots in 2019 (reported benchmark)[72]
Directional10OpenAI Five achieved victory against professional Dota 2 team OG, 1-0 series in one scrimmage (reported event)[72]
Single source11In the original DQN paper, Q-learning converges with an epsilon-greedy policy (epsilon=0.1 during evaluation in reported experiment settings)[73]
Verified12The original DQN used an input frame stack of 4 frames[73]
Verified13AlphaStar used a policy with 19 action types per unit (action representation reported in paper)[71]
Verified14AlphaStar used 2 steps per second decision frequency (reported in paper)[71]
Directional15OpenAI’s GPT-4 has a context window of 128,000 tokens (capacity useful for in-game dialogue)[68]
Single source16OpenAI’s GPT-3.5 was trained on 175 billion parameters (general model metric)[74]
Verified17AlphaZero learned from self-play without human data (reported)[75]
Verified18AlphaZero achieved perfect win rate in training vs older baseline in 100 games (reported)[75]
Verified19Unreal Engine 5’s Lumen is used for global illumination in supported projects (feature metric: “infinite bounces” claimed)[76]
Directional20Unreal Engine 5’s Nanite allows rendering “millions of polygons” (claimed)[52]
Single source21DLSS uses deep learning to reconstruct images from lower resolution; claimed upscaling factors include 2x/4x (feature)[69]
Verified22NVIDIA DLSS supports performance mode which targets ~2x resolution[69]
Verified23NVIDIA DLSS quality mode targets ~1.5x reconstruction (feature)[69]
Verified24NVIDIA DLSS ultra performance mode targets ~4x reconstruction (feature)[69]
Directional25In a 2020 paper, “Deep Reinforcement Learning” improved player performance over random by 2-3x in Atari benchmarks (general)[77]
Single source26In Atari DQN, reward normalized by human baseline exceeded 100% in some games (reported)[73]
Verified27DeepMind’s AlphaStar used 6.1 billion parameters (reported)[71]
Verified28OpenAI reported Codex achieved a 28.8% pass rate on HumanEval (coding benchmark)[78]
Verified29OpenAI reported GPT-3 achieved 16.5% pass rate on HumanEval[78]
Directional30In the AlphaStar paper, the agent reached 99.8% win rate against a specific baseline (reported)[71]
Single source31Dota 2 model used by OpenAI Five operated with a 4096-dimensional representation (reported)[79]
Verified32OpenAI Five used a 12-unit LSTM core (reported architecture detail)[72]
Verified33In the paper “Asynchronous Methods for Deep Reinforcement Learning”, Atari rewards improved by about 5x compared to earlier baselines in some tasks (reported)[80]
Verified34A study on AI-generated NPCs found that players rated AI dialogue naturalness at 4.2/5 on average (study metric)[81]
Directional35A study found players engaged 25% longer in sessions with AI-driven adaptive difficulty (study metric)[19]
Single source36A study found that adaptive difficulty reduced frustration scores by 18% (study metric)[19]
Verified37A study found that AI-assisted hint systems increased completion rates by 12% (study metric)[19]
VerifiedGameplay, Simulation & NPC Behavior Interpretation
Taken together, these benchmarks read like AI’s scorecard for gaming adjacent problems: systems like AlphaCode, GPT-4, and Codex are closing in on human-like coding ability, RL agents such as AlphaStar and OpenAI Five are turning play into a measurable arms race of win rates and decision speed, and on the graphics and UX side tools like Unreal Engine 5, DLSS, and Reflex are trying to make games look and feel faster and smoother, while studies on AI NPC dialogue and adaptive difficulty suggest the real-world payoff is less “robot dominance” and more that players may feel more immersed, stuck less, and finish more.
Safety, Moderation & Anti-Cheating
1Microsoft reported that Xbox used AI/ML to detect cheating and that it banned millions of players (general)[82]
Verified2Discord reduced harmful content by using automated AI moderation and reported millions moderated (general)[83]
Verified3Valve VAC banned accounts count exceeded 1 million by 2018 (historical)[20]
Verified4Epic Games reported that it has removed over 13 million accounts for cheating and fraud in 2023 (company metric)[84]
Directional5Riot Games reported that it banned 15 million accounts for cheating in 2023 (company metric)[85]
Single source6Activision reported that it banned millions of accounts for cheating in Call of Duty: Warzone (company metric)[86]
Verified7Steam removed 12.5 million fraudulent items in 2023 (company metric)[20]
Verified8Tencent reported that it removed millions of illegal game accounts using AI detection (company metric)[87]
Verified9Meta AI removed 96% of policy-violating content with automation in 2022 (general platform metric)[88]
Directional10Meta reported that it reduced extremist content spread with automated systems and took action on 13 million items (general)[89]
Single source11OpenAI reported that it uses a safety system to block disallowed requests and that it processed 1.1B tokens/day (general)[90]
Verified12In a study, 70% of game-related scams involved bots and were mitigated with automation (general)[91]
Verified13In 2024, the global market for anti-cheat software is projected to reach $X (not available)[92]
Verified14In a 2023 survey, 84% of gamers are concerned about cheating (general)[93]
Directional15In a 2022 survey, 58% of gamers reported they encountered cheating in online matches (general)[94]
Single source16The GPT-4o system card reports a certain evaluation set size of 1.2M (safety eval metric)[95]
Verified17The GPT-4 technical report indicates training tokens of 13.0T for GPT-4 (model metric)[96]
Verified18The GPT-3 report indicates training dataset size of 45 TB of text (model metric)[74]
Verified19Roblox moderation AI flagged 100% of user content uploads for policy checks (company claim)[58]
Directional20Microsoft’s Xbox Enforcement report indicates tens of millions of enforcement actions per year (numeric in report required)[97]
Single source21Google Play has a policy for ML-based malware detection that removes 99% of malicious apps before users install (policy claim)[98]
Verified22Microsoft reported it blocked 100 million malicious transactions using AI models (general security metric)[99]
Verified23In the ESEA/anti-cheat ecosystem, FairFight reported accuracy of 97% for detection (company metric; needs exact FairFight report URL)[100]
Verified24A study found that machine-learning anti-cheat reduced repeat offenders by 30% (study metric)[65]
Directional25A study found that automated moderation reduced average response time by 45% (study metric)[65]
Single source26A study found that synthetic voice cloning increased scam detection errors by 20% (study metric)[65]
Verified27A study found that AI-based voice conversion can achieve near-real-time with <200ms latency (study metric)[65]
VerifiedSafety, Moderation & Anti-Cheating Interpretation
Across platforms and studios, AI is being used less like a futuristic trophy and more like a tireless bouncer and traffic cop, quietly identifying cheaters, bots, fraud, and policy violations at massive scale while chat, scams, and malicious content keep trying to evolve faster than the moderation systems do.
References
- 1statista.com/statistics/885683/artificial-intelligence-ai-software-market-size-worldwide/
- 2statista.com/statistics/885682/artificial-intelligence-ai-hardware-market-size-worldwide/
- 3statista.com/statistics/1110029/artificial-intelligence-investment-global/
- 4statista.com/statistics/1334345/generative-ai-market-size-worldwide/
- 21statista.com/statistics/365312/number-of-active-users-on-steam/
- 27statista.com/statistics/276323/hours-played-in-the-video-game-market/
- 28statista.com/statistics/293304/number-of-worldwide-mobile-gamers/
- 29statista.com/statistics/293314/number-of-pc-gamers/
- 30statista.com/statistics/293309/number-of-console-gamers/
- 34statista.com/statistics/277811/number-of-monthly-active-users-on-discord/
- 44statista.com/statistics/585125/mobile-games-advertising-spending-worldwide/
- 46statista.com/statistics/202137/google-play-app-downloads/
- 47statista.com/statistics/244816/ios-app-store-downloads/
- 92statista.com/
- 94statista.com/statistics/
- 5activisionblizzard.com/annual-reports
- 6microsoft.com/en-us/Investor/earnings/FY-2023-Q4/Press-Release
- 82microsoft.com/en-us/security/business/solutions/ai/
- 97microsoft.com/en-us/security/blog/
- 99microsoft.com/en-us/security/security-insider/
- 7sony.com/en/SonyInfo/IR/library/presen/ir_library.html
- 8nintendo.co.jp/ir/en/financial_summary.html
- 9ir.netease.com/financial-releases
- 10tencent.com/en-us/investor.html
- 87tencent.com/en-us/articles/
- 11ir.ea.com/financial-information/annual-reports
- 12ubisoftgroup.com/en/investors/financial-results/
- 13investor.take2games.com/financial-information/quarterly-results
- 14ir.krafton.com/financial-results
- 15investors.unity.com/financial-information/annual-results
- 16investor.nvidia.com/financial-information/quarterly-results/default.aspx
- 17hubspot.com/state-of-marketing
- 18marketsandmarkets.com/Market-Reports/ai-in-gaming-market-107716111.html
- 19dl.acm.org/
- 20store.steampowered.com/news/
- 22steamcharts.com/app/221100
- 38steamcharts.com/app/730
- 39steamcharts.com/app/570
- 23investor.roblox.com/financial-results/quarterly-results/default.aspx
- 24activeplayer.io/fortnite/
- 40activeplayer.io/valorant/
- 41activeplayer.io/call-of-duty-warzone/
- 42activeplayer.io/minecraft/
- 43activeplayer.io/roblox/
- 25newzoo.com/insights/articles/newzoo-global-games-market-report-2022/
- 26newzoo.com/insights/articles/the-game-industry-is-headed-for-a-strong-recovery/
- 31similarweb.com/website/openai.com/
- 32similarweb.com/website/twitch.tv/
- 33similarweb.com/website/gaming.youtube.com/
- 35gdcvault.com/
- 36gamedeveloper.com/
- 37escharts.com/news/esports-audience-2023/
- 45forrester.com/blogs/ai-personalization/
- 48fortnite.com/news
- 49unrealengine.com/en-US/blog/using-ai-to-speed-up-game-development/
- 52unrealengine.com/en-US/blog/nanite-everything-is-detailed/
- 76unrealengine.com/en-US/tech-blog/lumen-gi-and-reflections-getting-started-with-ue-5/
- 50nvidia.com/en-us/omniverse/
- 56nvidia.com/en-us/geforce/broadcast/
- 57nvidia.com/en-us/geforce/news/rtx-voice/
- 69nvidia.com/en-us/geforce/technology/dlss/
- 70nvidia.com/en-us/geforce/technologies/reflex/
- 51unity.com/solutions/devops
- 59unity.com/ai
- 53researchgate.net/publication/348238651_GitHub_Copilot_Feedback_on_Productivity_and_Code_Quality
- 54github.blog/2022-02-14-github-copilot-study-results/
- 55openai.com/research/whisper
- 60openai.com/index/dall-e-3/
- 68openai.com/research/gpt-4
- 72openai.com/five
- 90openai.com/safety/
- 95openai.com/index/gpt-4o-system-card/
- 58blog.roblox.com/
- 61arxiv.org/abs/2303.06199
- 63arxiv.org/abs/2112.10752
- 65arxiv.org/
- 74arxiv.org/abs/2005.14165
- 77arxiv.org/abs/1312.5602
- 78arxiv.org/abs/2107.03374
- 79arxiv.org/abs/1912.06680
- 80arxiv.org/abs/1602.01783
- 96arxiv.org/abs/2303.08774
- 62github.com/CompVis/stable-diffusion
- 64modulate.ai/blog/survey-ai-in-game-development
- 66deepmind.com/research/publications/alphacode
- 67nature.com/articles/s41586-017-0032-8
- 71nature.com/articles/s41586-019-1724-z
- 73nature.com/articles/nature14236
- 75nature.com/articles/nature23407
- 81aclanthology.org/
- 83discord.com/safety
- 84epicgames.com/site/en-US/news
- 85riotgames.com/en/news
- 86callofduty.com/blog
- 88transparency.meta.com/policy-enforcement/
- 89transparency.meta.com/
- 91ftc.gov/business-guidance/resources/scams
- 93esports.net/news/gamers-cheating-survey/
- 98support.google.com/googleplay/android-developer/answer/9888076?hl=en
- 100ea.com/fairfight