Redigerer Generativ kunstig intelligens (avsnitt)

=== ''Deepfakes'' ===
''Deepfakes'' (engelsk, et [[teleskopord]] fra ''deep learning'', «[[dyp læring]]», og ''fake'', «falsk»<ref name="FoxNews2018">{{Cite news|last=Brandon|first=John|date=2018-02-16|title=Terrifying high-tech porn: Creepy 'deepfake' videos are on the rise|language=en-US|work=Fox News|url=http://www.foxnews.com/tech/2018/02/16/terrifying-high-tech-porn-creepy-deepfake-videos-are-on-rise.html|url-status=live|access-date=2018-02-20|archive-url=https://web.archive.org/web/20180615160819/http://www.foxnews.com/tech/2018/02/16/terrifying-high-tech-porn-creepy-deepfake-videos-are-on-rise.html|archive-date=2018-06-15}}</ref>) er KI-genererte filer som tar en person i et eksisterende bilde eller video og erstatte vedkommende med noen andres utseende ved bruk av [[Kunstig nevralt nettverk|kunstige nevrale nettverk]].<ref name=":3">{{cite web|last=Cole|first=Samantha|date=24. jan. 2018|title=We Are Truly Fucked: Everyone Is Making AI-Generated Fake Porn Now|url=https://www.vice.com/en_us/article/bjye8a/reddit-fake-porn-app-daisy-ridley|url-status=live|archive-url=https://web.archive.org/web/20190907194524/https://www.vice.com/en_us/article/bjye8a/reddit-fake-porn-app-daisy-ridley|archive-date=7. sep. 2019|access-date=4. mai 2019|website=Vice}}</ref> Deepfakes har mottatt utbredt oppmerksomhet og bekymringer for [[Pornografisk film|pornografiske videoer]] av [[Kjendis|kjendiser]], [[hevnporno]], falske nyheter, [[Bløff|bløffer]], [[Desinformasjon|helsedesinformasjon]] og finanssvindel.<ref name="HighSnobiety20182">{{Cite news|date=2018-02-20|title=What Are Deepfakes & Why the Future of Porn is Terrifying|language=en-US|work=Highsnobiety|url=https://www.highsnobiety.com/p/what-are-deepfakes-ai-porn/|url-status=live|access-date=2018-02-20|archive-url=https://web.archive.org/web/20210714032914/https://www.highsnobiety.com/p/what-are-deepfakes-ai-porn/|archive-date=2021-07-14}}</ref><ref>{{cite web|title=Experts fear face swapping tech could start an international showdown|url=https://theoutline.com/post/3179/deepfake-videos-are-freaking-experts-out|url-status=live|archive-url=https://web.archive.org/web/20200116140157/https://theoutline.com/post/3179/deepfake-videos-are-freaking-experts-out|archive-date=2020-01-16|access-date=2018-02-28|website=The Outline|language=en}}</ref><ref>{{Cite news|last=Roose|first=Kevin|date=2018-03-04|title=Here Come the Fake Videos, Too|language=en-US|work=The New York Times|url=https://www.nytimes.com/2018/03/04/technology/fake-videos-deepfakes.html|url-status=live|access-date=2018-03-24|archive-url=https://web.archive.org/web/20190618203019/https://www.nytimes.com/2018/03/04/technology/fake-videos-deepfakes.html|archive-date=2019-06-18|issn=0362-4331}}</ref><ref>{{cite arXiv|eprint=1910.03810|class=cs.LG|first1=Marco|last1=Schreyer|first2=Timur|last2=Sattarov|title=Adversarial Learning of Deepfakes in Accounting|language=en|last3=Reimer|first3=Bernd|last4=Borth|first4=Damian|year=2019}}</ref><ref>{{Cite journal|last1=Menz|first1=Bradley|title=Health Disinformation Use Case Highlighting the Urgent Need for Artificial Intelligence Vigilance|date=2024|url=https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2811333|access-date=4 February 2024|journal=JAMA Internal Medicine|volume=184|issue=1|pages=92–96|doi=10.1001/jamainternmed.2023.5947|pmid=37955873|s2cid=265148637|archive-date=February 4, 2024|archive-url=https://web.archive.org/web/20240204083644/https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2811333|url-status=live}}</ref> Innspill fra både industrier og myndigheter om å oppgave og begrense bruken har oppstått som følge av dette.<ref name=":21">{{Cite web|title=Join the Deepfake Detection Challenge (DFDC)|url=https://deepfakedetectionchallenge.ai/|url-status=live|archive-url=https://web.archive.org/web/20200112102819/https://deepfakedetectionchallenge.ai/|archive-date=2020-01-12|access-date=2019-11-08|website=deepfakedetectionchallenge.ai}}</ref><ref name=":5">{{Cite web|last=Clarke|first=Yvette D.|date=2019-06-28|title=H.R.3230 – 116th Congress (2019-2020): Defending Each and Every Person from False Appearances by Keeping Exploitation Subject to Accountability Act of 2019|url=https://www.congress.gov/bill/116th-congress/house-bill/3230|url-status=live|archive-url=https://web.archive.org/web/20191217110329/https://www.congress.gov/bill/116th-congress/house-bill/3230|archive-date=2019-12-17|access-date=2019-10-16|website=www.congress.gov}}</ref>

[[Logically]], et selskap som driver med faktasjekk, fant i juli 2023 at populære generative KI-modeller som Midjourney, DALL-E 2 og Stable Diffusion produserte mulige desinformasjonsbilder når de ble kommandert til å gjøre det, som for eksempel bilder av [[valgfusk]] i USA og [[Muslim|muslimske]] kvinner som støtter [[Bharatiya Janata Party]], et [[Hinduistisk nasjonalisme|hinduistisk nasjonalistisk]] parti i [[India]].<ref>{{Cite web|date=27. jul. 2023|title=New Research Reveals Scale of Threat Posed by AI-generated Images on 2024 Elections|url=https://www.logically.ai/press/new-research-reveals-scale-of-threat-posed-by-ai-generated-images-on-2024-elections|url-status=dead|archive-url=https://web.archive.org/web/20231003172151/https://www.logically.ai/press/new-research-reveals-scale-of-threat-posed-by-ai-generated-images-on-2024-elections|archive-date=3. okt. 2023|access-date=6. jul. 2024|website=Logically}}</ref><ref>{{Cite web|last=Lawton|first=Graham|date=12. sep. 2023|title=Disinformation wars: The fight against fake news in the age of AI|url=https://www.newscientist.com/article/mg25934563-200-disinformation-wars-the-fight-against-fake-news-in-the-age-of-ai/|access-date=2024-07-05|website=[[New Scientist]]|language=en-US}}</ref>