AI倫理について

AI倫理を巡る各種議論や実際の事例について国際基督教大学で講義した際のスライドです

Transcript

1. "*ྙཧʹ͍ͭͯ 2024.5.29 ߹ಉձࣾpratya ஑ా ହ߂
2. ࣗݾ঺հ ஑ా ହ߂ʢIKEDA Yasuhiroʣ • 2010೥ɺNTTݚڀॴʹೖࣾɻωοτϫʔΫͷ҆શͳӡ༻ʹ޲͚ͯɺػցֶशΛ׆༻ͨ͠ ҟৗݕ஌ΞϧΰϦζϜΛ։ൃɻޙʹʮAIҟৗ༧ஹݕ஌ιϦϡʔγϣϯ ˏDeAnoSʯ ͱͯ͠NTTΞυόϯεςΫϊϩδΑΓ੡඼Խ • 2019೥ɺPKSHA TechnologyʹࢀըɻΤϯδχΞϦϯάϚωʔδϟʔͱͯ͠ɺઌਐతͳ ٕज़ݕ౼͔Βࣄۀಋೖʹ޲͚ͨ։ൃ·ͰΛखֻ͚ΔɻΫϨδοτΧʔυෆਖ਼ݕ஌Ξϧΰ ϦζϜͷ։ൃ΍σΟʔϓϥʔχϯάΛ༻͍ͨΦʔσΟΤϯε֦ுɺ޿ࠂίϯόʔδϣϯ ༧ଌϞσϧߏங౳ͷ࣮੷͋Γ • 2021೥ɺσʔλαΠΤϯεʹΑΔࣾձ՝୊ղܾΛՃ଎͢Δ΂͘߹ಉձࣾpratyaΛઃཱ 2
3. Ξ΢τϥΠϯ 1. AIٕज़ͷࣾձ࣮૷ͱྙཧత໰୊ʹ͍ͭͯ 2. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎલʣ 3. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎ߱ʣ 4. AIྙཧΛ८Δ࣮ࡍͷࣄྫ 5. ·ͱΊ 3
4. "*ࢢ৔ن໛ͷਪҠ • ૯຿লͷௐࠪʹΑΔͱɺੈքͷAIࢢ৔ن໛͸2022೥࣌఺Ͱ໿18ஹԁͱͳ͓ͬͯΓ 2030೥·Ͱ͸Ճ଎౓తʹ੒௕͢Δͱ͞Ε͍ͯΔ • ੜ੒AIͷ୆಄ʹΑͬͯɺࢢ৔ͷ੒௕͸͞ΒͳΔՃ଎͕ݟࠐ·Ε͍ͯΔ Ø 2023೥1݄ʹϦϦʔε͞ΕͨAzure OpenAIͷಋೖࣾ਺͸1೥Ͱ2,300ࣾʹ 4 Artificial intelligence (AI) market size worldwide in 2021 with a forecast until 2030 https://www.statista.com/statistics/1365145/artificial-intelligence-market-size/ 日本マイクロソフトが生成AIサービスの利用社数を公開し続ける理由 https://japan.zdnet.com/article/35213126/
5. ྙཧత໰୊ͷࣄྫ • ब৬৘ใαΠτʮϦΫφϏʯͷؔ࿈αʔϏεʮϦΫφϏDMPϑΥϩʔʯʹ͓͍ͯɺ ొ࿥Ϣʔβʔͷֶੜ͕اۀͷ಺ఆΛࣙୀ͢Δ֬཰ΛAIʹΑͬͯ༧ଌɻσʔλ͸࠾༻اۀ ΁ൢച͞Εͨ • 2019೥8݄ʹ੓෎ͷݸਓ৘ใอޢҕһձ͔ΒקࠂΛड͚ɺαʔϏε͸ഇࢭʹɻ9݄ʹ͸ ްੜ࿑ಇলʹΑΔߦ੓ࢦಋ͕࣮ࢪ͞Εͨ →AI׆༻ʹ͓͚Δྙཧత໰୊ͷऔΓѻ͍ʹ஫໨͕ू·Δ 5 日本経済新聞「リクナビに行政指導 辞退率販売、職安法に違反」 https://www.nikkei.com/article/DGXMZO49487690W9A900C1MM0000/
6. Ξ΢τϥΠϯ 1. AIٕज़ͷࣾձ࣮૷ͱྙཧత໰୊ʹ͍ͭͯ 2. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎલʣ 3. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎ߱ʣ 4. AIྙཧΛ८Δ࣮ࡍͷࣄྫ 5. ·ͱΊ 6
7. +PCJO
   "OOB
   .BSDFMMP
   *FODB
   BOE
   &GGZ
   7BZFOB
   5IF
   HMPCBM
   MBOETDBQF
   PG
   "*
   FUIJDT
   HVJEFMJOFT /BUVSF
   .BDIJOF
   *OUFMMJHFODF 
    
   
   • աڈ5೥ʹ౉Δɺ͋ΒΏΔࠃͷاۀɾݚڀॴɾ࣏ࣗମͷൃදͨ͠AIྙཧΨΠυϥΠϯͷ಺༰Λௐࠪ • සग़͢ΔΩʔϫʔυΛ෼ੳͨ݁͠Ռʮಁ໌ੑʢઆ໌ੑɺղऍੑʣʯʮਖ਼ٛɾެฏੑʯʮඇ༗֐ੑʯ ʮ੹೚ੑʯʮϓϥΠόγʔʯͷ5߲໨͕ɺա൒਺ͷΨΠυϥΠϯ಺Ͱग़ݱ 7 https://www.nature.com/articles/s42256-019-0088-2
8. ಁ໌ੑʢઆ໌ੑɾղऍੑʣ 8 • AIͷ൑அʹ͍ͭͯઆ໌ɺղऍͰ͖Δ͜ͱ ü AIʹΑΔة֐ͷ཈੍ ü AIʹΑΔޮՌͷ૿෯ ü ๏తͳཧ༝ ü AIͱͷ৴པؔ܎ߏங • AIͷಁ໌ੑ޲্ʹ޲͚ͯɺԼه͕ॏཁͱߟ͑ΒΕ͍ͯΔ ü AIγεςϜ։ൃʹ͓͚Δ৘ใͷ։ࣔ ü ιʔείʔυɺσʔλͷ։ࣔ ü ؂ࠪੑ
9. ਖ਼ٛɾެฏੑ 9 • ʮਖ਼ٛʯͱ͸… ü ެฏੑɺόΠΞεɾࠩผͷ཈ࢭɺଟ༷ੑɺΠϯΫϧʔδϣϯɺAIͷܾఆʹର͢Δҟٞਃཱͯ͠ɾ ٹࡁાஔɺAI΍σʔλɾAIͷԸܙ΁ͷΞΫηεʹର͢Δެฏੑɺ࿑ಇࢢ৔ɾຽओओٛɾࣾձ໰୊ ΁ͷߩݙɺֶशσʔλͷଟ༷ੑ୲อɺetc. • ʮਖ਼ٛʯΛ୲อ͢ΔͨΊʹ͸ҎԼ͕ॏཁͰ͋Δͱߟ͑ΒΕ͍ͯΔ ü ن֨ͷࡦఆ΍໌ࣔతͳنൣͷϞσϧ૊ΈࠐΈ ü ಁ໌ੑΛ୲อ͠ɺݖར΍੍ݶʹ͍ͭͯެ։͢Δ ü ςετɾ؂ࢹɾ؂ࠪͷపఈ ü ๏ʹΑΔن੍ͷൃలɾڧԽɺҟٞਃཱͯ͠ͷݖརɾٹࡁાஔͷ֬อ ü ֤छ੍౓ͷΞοϓσʔτ ü ੓࣏తΞΫγϣϯɺֶࡍతɾଟ༷ͳ࿑ಇ؀ڥɺΠϯΫϧʔγϒͳࣾձɺརӹͷ෼഑ɺetc.
10. .FISBCJ
    /JOBSFI
    FU
    BM
    "
    TVSWFZ
    PO
    CJBT
    BOE
    GBJSOFTT
    JO
    NBDIJOF
    MFBSOJOH "$.
    $PNQVUJOH
    4VSWFZT
    $463 
     
     10 • طଘݚڀΛௐࠪ͠ɺػցֶशʹ͓͚Δެฏੑʹ͍֤ͭͯछఆٛΛ෼ྨ ü 23छͷόΠΞεʢྺ࢙తόΠΞεɺਓޱόΠΞεɺetc.ʣ ü 6छͷࠩผʢ௚઀తɾඇ௚઀తࠩผɺ੍౓ʹΑΔࠩผɺetc.ʣ ü 10छͷެฏੑʢ౳ՁΦοζɺػձۉ౳ɺແҙࣝʹΑΔެฏɺetc.ʣ • "ެฏͳػցֶश"ͷख๏ΛʮલॲཧʯʮதॲཧʯʮޙॲཧʯͰ෼ྨ ü લॲཧɿֶशσʔλʹ಺ࡏ͢ΔࠩผΛࣄલʹऔΓআ͘ ü தॲཧɿطଘͷֶशϞσϧʹରͯ͠ɺֶशͷաఔʹ͓͍ͯࠩผ͕ى͜Βͳ͍Α͏ ໨తม਺΍੍໿৚݅ΛՃ͑Δ ü ޙॲཧɿ௨ৗ௨ΓʹϞσϧΛֶशͨ͠ޙͰɺ༧ଌ݁ՌΛެฏੑ͕ຬͨ͞ΕΔΑ͏ิਖ਼
11. ެฏੑͷτϨʔυΦϑͱྙཧֶ 11 • ެฏੑͷई౓ʹ͸ʮଐੑؒͰͷ֨ࠩʢImpact Disparityʣͷճආʯͱ ʮଐੑͷΑ͏ͳηϯγςΟϒͳ৘ใΛ൑அʹ༻͍ΔʢTreatment Disparityʣ͜ͱͷճආʯ ͷ2छ͕ଘࡏ* ü લऀͷྫɿ༩৴ͷείΞʹ͍ͭͯɺੑผؒͰείΞͷ෼෍͕౳͘͠ͳΔΑ͏ʹ͢Δ ʢެฏੑʹ͓͚Δ”Demographic Parity”ʣ ü ޙऀͷྫɿ༩৴Λ͓͜ͳ͏ϞσϧͷΠϯϓοτʹੑผͷ৘ใΛ༻͍ͳ͍ ʢެฏੑʹ͓͚Δ”Fairness Through Unawareness”ʣ • ྙཧֶͷ؍఺ͰɺલऀΛʮ݁Ռʯʹண໨ͨ͠ޭརओٛతΞϓϩʔνɺޙऀΛʮҙਤʯʹ ண໨ͨٛ͠຿࿦తΞϓϩʔνͱղऍ͢Δݚڀ΋͋Δ** ** Mougan, Carlos, and Joshua Brand. "Kantian Deontology Meets AI Alignment: Towards Morally Robust Fairness Metrics." arXiv preprint arXiv:2311.05227 (2023). https://arxiv.org/abs/2311.05227 * Zafar, Muhammad Bilal, et al. "Fairness constraints: Mechanisms for fair classification." Artificial intelligence and statistics. PMLR, 2017. https://proceedings.mlr.press/v54/zafar17a.html
12. Ξ΢τϥΠϯ 1. AIٕज़ͷࣾձ࣮૷ͱྙཧత໰୊ʹ͍ͭͯ 2. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎલʣ 3. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎ߱ʣ 4. AIྙཧΛ८Δ࣮ࡍͷࣄྫ 5. ·ͱΊ 12
13. $IBU(15ͷग़ݱͱੜ੒"*ͷོ੝ • 2023೥3݄ʹOpenAI͕ࣾର࿩ܕAIʮChatGPTʢGPT-4ʣʯΛϦϦʔεɻਓؒͷΑ͏ͳࣗવͳର࿩ ͕Ͱ͖Δ͜ͱͰੈքதͰେ͖ͳ஫໨ΛूΊͨ • ର࿩ܕAIͱڝ͏Α͏ʹͯ͠ɺಈը૾ͷੜ੒AIʹ͍ͭͯ΋։ൃ͕੝ΜʹɻStable Diffusion, Midjourney, SoraͳͲ༷ʑͳϞσϧ͕ϦϦʔε͞Εͨ 13 Sora https://openai.com/sora ChatGPT https://chat.openai.com/
14. $IBU(15ͷٕज़ཁૉɿ5SBOTGPSNFS • 2017೥ʹGoogle BrainͷݚڀऀΒʹΑΔ࿦จ”Attention Is All You Need”ͰఏҊ͞Εͨख๏ɻ຋༁౳ ͷλεΫʹ͓͍ͯେ෯ͳਫ਼౓޲্Λ࣮ݱ ü GPTʢGenerative Pre-trained TransformerʣͰ͸Decoder෦෼ͷΈΛ૊Έ߹Θͤͯ࢖༻ • 2020೥ʹɺOpenAIͷݚڀऀΒʹΑΔ࿦จ”Scaling Laws for Neural Language Models”Ͱɺ Transformerͷਫ਼౓͕ϞσϧαΠζʗσʔληοταΠζʗܭࢉྔ͕૿͑Δʹ࿈Εͯ΂͖৐ଇʹ ैͬͯ޲্͠ɺͦͷ্ݶʹ͍ͭͯ͸؍ଌ͞Εͳ͔ͬͨ͜ͱ͕ࣔ͞Εͨ →ϦιʔεΛ౤͡Δ΄Ͳਫ਼౓্͕͕Γଓ͚ΔՄೳੑΛࣔࠦ 14 左図：Attention Is All You Need https://arxiv.org/abs/1706.03762 右図：Scaling Laws for Neural Language Models https://arxiv.org/abs/2001.08361
15. $IBU(15ͷٕज़ཁૉɿ3-)' • ChatGPTͰ͸Ϟσϧͷେن໛ԽʹՃ͑ͯɺਓखʹΑΔϥϕϦϯάΛ׆༻ͨ͠ڧԽֶश ʢRLHFɿReinforcement Learning from Human FeedbackʣʹΑͬͯɺΑΓਓؒʹͱͬͯ޷·͍͠ ର࿩Λ͓͜ͳ͏Ϟσϧ΁ͱνϡʔχϯά Ø Step 1ɿਓखͰ࡞੒͞Εͨʮ޷·͍͠ճ౴ʯΛਖ਼ղσʔλͱͨ͠ڭࢣ͋ΓֶशʹΑΔϑΝΠϯνϡʔχϯά Ø Step 2ɿೖྗจʹର͢Δෳ਺ͷճ౴ީิʹରͯ͠ɺਓखʹΑΔϥϯΫ෇͚Λ࣮ࢪɻͦͷσʔλΛ༻͍ͯɺೖྗจ ʹର͢Δճ౴จͷϥϯΫΛ༧ଌ͢ΔReward ModelʢRMʣΛֶश Ø Step 3ɿRMʹΑΔධՁ͕࠷େͱͳΔΑ͏ͳग़ྗΛ͓͜ͳ͏ϞσϧΛڧԽֶशͰ࡞੒ 15 Training language models to follow instructions with human feedback https://arxiv.org/pdf/2203.02155
16. ը૾ੜ੒"*ͷٕज़ཁૉʢҰྫʣ • Stable Diffusionͷߏ੒ཁૉ Ø Variational AutoencoderɿσΟʔϓϥʔχϯάʹΑΔσʔλѹॖͷٕज़ɻը૾σʔλͷѹॖʹ࢖༻ Ø Diffusion ModelɿϊΠζ͔Βը૾Λੜ੒͢ΔϓϩηεΛֶश͢ΔϞσϧɻֶशσʔλʹରͯ͠෇༩ͨ͠ϊΠζΛ আڈ͢ΔΑ͏ͳϞσϧΛֶशͤ͞Δ Ø Denoising U-NetɿDiffusion Modelʹ͓͍ͯϊΠζআڈʹ༻͍ΒΕΔσΟʔϓϥʔχϯάͷϞσϧ Ø CLIPɿը૾ͱςΩετΛಉ͡જࡏۭؒʹຒΊࠐΉϞσϧɻςΩετͷຒΊࠐΈʹ͸Transformer͕༻͍ΒΕ͍ͯΔ 16 High-Resolution Image Synthesis with Latent Diffusion Models https://arxiv.org/pdf/2112.10752 "a photograph of an astronaut riding a horse" Wikipedia : Stable Diffusion https://ja.wikipedia.org/wiki/Stable_Diffusion
17. 0QFO"*
    l(15
    5FDIOJDBM
    3FQPSUz
    BS9JW QSFQSJOU
    BS9JW   • GPT-4ͷొ৔ʹΑͬͯAI͕༷ʑͳ࣭໰ʹ౴͑ΒΕΔΑ͏ʹͳͬͨ͜ͱͰɺྙཧతͳ໰୊ʹؔ͢Δ஫໨ ͕ߴ·Δ • OpenAI͕GPT-4ͷϦϦʔεͱಉ࣌ʹެ։ͨ͠Technical ReportͰ͸ɺର࿩ܕAI๊͕͑Δةݥੑͱͦͷ ରॲʹ͍ͭͯղઆɻةݥੑͷྫͱͯ͠ҎԼΛڍ͍͛ͯͨ Ø ϋϧγωʔγϣϯʢݬ֮ɺͰͬͪ͋͛ʣɺϓϥΠόγʔɺ༗֐ͳ಺༰ɺ৘ใૢ࡞ɺةݥͳ૑ൃతߦಈɺetc. 17 GPT-4 Technical Report https://arxiv.org/abs/2303.08774
18. .PUPLJ
    'BCJP
    7BMEFNBS
    1JOIP /FUP
    BOE
    7JDUPS
    3PESJHVFT
    .PSF
    IVNBO
    UIBO
    IVNBO
    .FBTVSJOH
    $IBU(15 QPMJUJDBM
    CJBT 1VCMJD
    $IPJDF 
     
     • ChatGPTͷ੓࣏తͳόΠΞεʹ͍ͭͯධՁͨ͠࿦จɻChatGPTʹରͯ͠ຽओౘʗڞ࿨ౘͦΕͧΕͷ ཱ৔ΛऔΒ্ͤͨͰͷճ౴ͱɺ੓࣏తཱ৔Λࢦࣔ͠ͳ͍σϑΥϧτͷճ౴Λൺֱͨ͠ͱ͜Ζɺ ChatGPT͸ຽओౘدΓʢࠨ೿دΓʣͷόΠΞε͕͋Δ͜ͱ͕֬ೝ͞Εͨ • ถࠃͱಉ༷ʹ੓࣏తʹೋۃԽ͍ͯ͠ΔϒϥδϧɾӳࠃΛର৅ͱͨ͠ධՁͰ΋ɺ͍ͣΕ΋ࠨ೿دΓͷ ݁Ռ͕ಘΒΕͨ • ֶशͷݩͱͳͬͨ΢Σϒσʔλࣗମ΍ɺֶश࣌ʹߦΘΕͨΫϦʔχϯά΍৘ใ௥Ճʹ͓͚ΔόΠΞε ͕Өڹ͍ͯ͠ΔՄೳੑ͕͋Δ 18 https://arxiv.org/abs/2303.08774
19. 4PNFQBMMJ
    (PXUIBNJ
    FU
    BM
    %JGGVTJPO
    BSU
    PS
    EJHJUBM
    GPSHFSZ
    JOWFTUJHBUJOH
    EBUB
    SFQMJDBUJPO
    JO
    EJGGVTJPO
    NPEFMT 1SPDFFEJOHT
    PG
    UIF
    *&&&$7'
    $POGFSFODF
    PO
    $PNQVUFS
    7JTJPO
    BOE
    1BUUFSO
    3FDPHOJUJPO
     • Diffusion Modelֶ͕शσʔλΛͦͷ··࠶ݱͨ͠σʔλΛग़ྗͯ͠͠·͏Մೳੑʹ͍ͭͯௐࠪ • ެ։σʔληοτLAIONΛ༻͍ͯධՁɻը૾ͷΩϟϓγϣϯΛؚΜͩϓϩϯϓτΛStable Diffusionʹ ༩͑Δ͜ͱͰɺର৅ͷը૾ʹࠅࣅͨ͠ग़ྗΛಘΔ͜ͱʹ੒ޭͨ͠ 19 https://arxiv.org/pdf/2212.03860
20. Ξ΢τϥΠϯ 1. AIٕज़ͷࣾձ࣮૷ͱྙཧత໰୊ʹ͍ͭͯ 2. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎલʣ 3. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎ߱ʣ 4. AIྙཧΛ८Δ࣮ࡍͷࣄྫ 5. ·ͱΊ 20
21. ஶ࡞ݖʹؔ͢Δ໰୊ɿχϡʔεهࣄͷֶश 21 • ChatGPT౳ͷֶशσʔλͱͯࣗࣾ͠ͷίϯςϯπΛ࢖༻͞Εͨͱͯ͠ɺχϡʔϤʔΫɾλΠϜζࢴ ͕OpenAIͱMicrosoftΛஶ࡞ݖ৵֐Ͱఏૌ • ૌঢ়Ͱ͸ɺGPT-4ʹಛఆͷϓϩϯϓτΛೖྗ͢ΔͱχϡʔϤʔΫɾλΠϜζࢴͷهࣄͱಉ͡಺༰͕ ؙ͝ͱग़ྗ͞Εͯ͠·͏ͱओுʢOpenAIଆ͸ʮϋοΩϯά͞Εͨʯͱ൓࿦ʣ The Times Sues OpenAI and Microsoft Over A.I. Use of Copyrighted Work https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html
22. ஶ࡞ݖʹؔ͢Δ໰୊ɿը૾ੜ੒Λ༻͍ͨάϥϏΞ 22 • ूӳࣾʹΑΓग़൛͞Εͨը૾ੜ੒AIΛ༻͍ͨάϥϏΞࣸਅू͕ɺSNS্Ͱͷ൷൑Λड͚ͯൢചதࢭʹ • ஶ࡞ݖΛ໰୊ࢹ͢Δ੠ʹՃ͑ɺάϥϏΞΛҾୀ࣮ͨ͠ࡏͷਓ෺ʹࣅ͍ͯΔͱ͍ͬͨࢦఠ΋ड͚͍ͯͨ ＡＩが生み出した女性モデルのグラビア写真集、集英社が販売中止…著作権を問題視する声 https://www.yomiuri.co.jp/culture/20230607-OYT1T50185/
23. ஶ࡞ݖ໰୊ʹର͢ΔऔΓ૊Έ 23 • จԽி͸༗ࣝऀΒͷٞ࿦΍ύϒϦοΫίϝϯτΛ౿·͑ͯɺ2݄ʹʮAIͱஶ࡞ݖʹؔ͢Δߟ͑ํʯͷ ҊΛެ։ Ø https://www.bunka.go.jp/seisaku/bunkashingikai/chosakuken/hoseido/r05_07/pdf/94011401_02.pdf • ݱߦͰ͸ݪଇͱͯ͠ೝΊΒΕ͍ͯΔɺஶ࡞෺ΛແஅͰ༻͍ͨAIֶशʹ͍ͭͯɺಛఆͷΫϦΤΠλʔͷ ࡞඼Λҙਤతʹֶशͤ͞Δߦҝ΍ɺݕࡧΛ੍ݶ͢ΔાஔΛߨͨ͡΢ΣϒαΠτͷσʔλͷֶश͸ஶ࡞ ݖͷ৵֐ʹ౰ͨΔՄೳੑ͕͋Δ఺ͳͲΛࣔͨ͠ 「AIと著作権に関する考え方」の案が示される 文化庁 https://www3.nhk.or.jp/news/html/20240229/k10014374871000.html
24. ޡ৘ใʹؔ͢Δ໰୊ɿ࣮ࡏ͠ͳ͍൑ྫͷҾ༻ 24 • χϡʔϤʔΫभͰߦΘΕͨຽࣄૌুʹ͓͍ͯɺหޢ͕࢜ࢿྉ࡞੒ʹChatGPTΛ༻͍ͨ͜ͱͰ ϋϧγωʔγϣϯʹΑͬͯੜ੒͞Ε࣮ͨࡏ͠ͳ͍൑ྫ͕Ҿ༻͞Ε͍ͯͨ͜ͱ͕ൃ֮ • ࣄ݅Λड͚ͯɺςΩαεभͷ͋Δ࿈๜൑ࣄ͸ʮఏग़ࢿྉͷ͍͔ͳΔ෦෼΋ੜ੒AIʹΑͬͯ࡞੒͞Ε ͍ͯͳ͍͜ͱΛূݴ͢Δ͔ɺੜ੒AIʹΑͬͯ࡞੒͞Εͨ෦෼͸ਖ਼֬ੑΛ֬ೝͰ͖ΔΑ͏໌֬Խ͢ Δʯ͜ͱΛ໋ྩͨ͠ Lawyer Used ChatGPT In Court—And Cited Fake Cases. A Judge Is Considering Sanctions https://www.forbes.com/sites/mollybohannon/2023/06/08/lawyer-used-chatgpt-in-court-and-cited-fake-cases-a-judge-is-considering-sanctions/?sh=3e8194d97c7f
25. ޡ৘ใͷ໰୊ʹର͢ΔऔΓ૊Έɿ0&$%ʹΑΔࠃࡍࢦ਑ʮ"*ݪଇʯ 25 • ܦࡁڠྗ։ൃػߏʢOECDʣ͕ࡦఆ͢ΔAIར༻ͷࠃࡍࢦ਑ʮAIݪଇʯʹ͍ͭͯɺੜ੒AIͷ୆಄Λ ౿·͑ͨվఆҊ͕5݄ʹఏࣔ͞Εͨ Ø https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449 • վఆҊͷதͰ͸ɺAI։ൃऀΒʹରِͯ͠৘ใɾޡ৘ใ΁ͷରॲΛٻΊΔ߲໨͕৽ͨʹ੝Γࠐ·Εͨ ＯＥＣＤが生成ＡＩ対策、開発者に偽・誤情報への対処求める…閣僚理事会で指針改定案採択へ https://www.yomiuri.co.jp/economy/20240430-OYT1T50220/
26. Ξϊςʔγϣϯʹ͓͚Δ࿑ಇࡡऔ໰୊ 26 • ChatGPT͕๫ྗతɾࠩผతͳग़ྗΛ͠ͳ͍ͨΊͷֶशσʔλ࡞੒ΛɺέχΞͷ࿑ಇऀΛޏ༻͢Δ Samaࣾʹґཔ • ࢠڙ΁ͷੑతٮ଴΍ࡴਓͳͲͷ಺༰ΛؚΉ๲େͳςΩετ΁ͷϥϕϧ෇͚Λɺ࣌څ2υϧఔ౓ͷ ௿௞ۚͰ࣮ࢪ • ΠϯλϏϡʔΛड͚ͨ4໊ͷ࿑ಇऀ͕͍ͣΕ΋ɺϝϯλϧϔϧεʹਂࠁͳѱӨڹΛड͚ͨͱड़΂ͨ Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic https://time.com/6247678/openai-chatgpt-kenya-workers/
27. ෆద੾ͳσʔλΛ༻͍ͨ"*ֶशͷ໰୊ 27 • ը૾ੜ੒AIͷֶश༻σʔληοτʮLAION 5Bʯʹ਺ඦຕͷࣇಐͷੑతٮ଴ը૾ؚ͕·Ε͍ͯͨ ͜ͱ͕ถελϯϑΥʔυେͷݚڀʹΑΓൃ֮ • LAION 5BΛֶशʹར༻͍ͯ͠ΔStability AIࣾ͸ϞσϧͷΞοϓσʔτ΍ϓϩϯϓτͷϑΟϧλϦϯά ʹΑͬͯ໰୊ʹରԠɻҰํɺಉ༷ʹLAION 5BΛར༻͍ͯ͠ΔMidjourney͔Βͷίϝϯτ͸ແ͔ͬͨ Stable Diffusion 1.5 Was Trained On Illegal Child Sexual Abuse Material, Stanford Study Says https://www.forbes.com/sites/alexandralevine/2023/12/20/stable-diffusion-child-sexual-abuse-material-stanford-internet-observatory/
28. Ξ΢τϥΠϯ 1. AIٕज़ͷࣾձ࣮૷ͱྙཧత໰୊ʹ͍ͭͯ 2. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎલʣ 3. AIྙཧʹؔ͢Δٞ࿦ʢChatGPTҎ߱ʣ 4. AIྙཧΛ८Δ࣮ࡍͷࣄྫ 5. ·ͱΊ 28
29. ·ͱΊ 29 • ੜ੒AIͷ୆಄ʹΑΓAIٕज़ͷࣾձ࣮૷͕Ճ଎͢ΔҰํͰɺྙཧ໘ʹؔ͢Δ՝୊͸਺ଟ͘ ੵΈ࢒͞Ε͓ͯΓɺ࣮ࣾձʹ͓͍ͯ΋طʹ༷ʑͳ໰୊͕ੜ͍ͯ͡Δ • ஶ࡞ݖอޢ΍ޡ৘ใରࡦͳͲҰ෦ͷ໰୊ʹ͍ͭͯ͸੍౓ԽͷऔΓ૊Έ͕ਐΉ΋ͷͷɺ ެฏੑͷτϨʔυΦϑ΍σʔλʹ಺ࡏ͢ΔόΠΞεͳͲɺຊ࣭తʹղܾͷ೉͍͠໰୊ͱ ͲͷΑ͏ʹରቂ͢Δ͔ʹ͍ͭͯ͸ࣾձతʹ߹ҙܗ੒ΛਐΊ͍ͯ͘ඞཁ͕͋ΓɺͦͷͨΊ ʹ͸֤ਓ͕৘ใྙཧʹؔ৺Λ࣋ͪɺ࠷৽ͷಈ޲Λ೺Ѳ͢Δ͜ͱ͕ॏཁͰ͋Δ