Andreea Lisievici

Ofcom Publishes Paper on Mitigating Deepfake Harms

AI, Deepfakes |

Ofcom has published a discussion paper on the impacts and mitigation of deepfakes, which are AI-generated or manipulated audio-visual content that misrepresents people or events. Deepfakes have caused harm to public figures and ordinary people, often through non-consensual sexual content, financial scams, or political disinformation. Ofcom’s recent poll reveals that nearly half of respondents aged 8-15 and 16+ have encountered deepfakes in the past six months. Under the Online Safety Act, online services must address harmful content, including deepfakes, through measures such as watermarking, detection classifiers, and media literacy efforts.

Ofcom Publishes Paper on Mitigating Deepfake Harms Read More »

FTC Warns Against Misleading Anonymization Claims Through Hashing

FTC, Hashing |

The Federal Trade Commission (FTC) emphasizes that hashing does not render data anonymous, cautioning companies against misleading privacy claims. Hashing, which transforms data like email addresses into consistent numerical values, may obscure the original data but still allows user identification. The FTC has taken action against companies such as Nomi and BetterHelp for improperly using hashed data, demonstrating that hashed identifiers can still cause privacy harms. The calls for companies adhere to truthful privacy practices, highlighting recent cases where user tracking persisted through unique identifiers.

FTC Warns Against Misleading Anonymization Claims Through Hashing Read More »

Google Halts Plans to Phase Out Third-Party Cookies

Cookies, Google |

Google has reversed its decision to phase out third-party cookies from its Chrome browser, opting instead to provide users with the option to block these cookies across their web browsing. This change, announced on 22 July 2024, is said to come after feedback from regulators and stakeholders, including the UK’s Competition and Markets Authority. Google says their new strategy aims to enhance user privacy while maintaining a vibrant ad-supported internet. However, concerns remain that this approach could still limit third-party advertising and shift ad spending towards Google’s search ads, raising potential antitrust issues.

Google Halts Plans to Phase Out Third-Party Cookies Read More »

FCC Proposes New AI Disclosure Rules for Political Ads

AI |

The FCC has introduced a proposal to standardize the disclosure of AI-generated content in political advertisements. This Notice of Proposed Rulemaking aims to bring uniformity to state laws regulating AI use in political campaigns. The proposal requires broadcasters, cable operators, Direct Broadcast Satellite (DBS) providers, and Satellite Digital Audio Radio Services (SDARS) licensees to disclose AI-generated content in political ads through on-air announcements and notices in political files. The initiative seeks to enhance transparency and help voters make informed decisions, amid concerns about AI-generated “deepfakes” misleading the public.

FCC Proposes New AI Disclosure Rules for Political Ads Read More »

NIST Released Guidance on Mitigating Generative AI Risks

GenAI, NIST |

NIST has introduced the AI RMF Generative AI Profile (NIST AI 600-1), offering a framework to help organizations manage risks specific to generative AI. This profile identifies 12 unique risks, such as increased cybersecurity threats, the spread of misinformation, and “hallucinations” in AI outputs. The document details over 200 actions developers can implement to mitigate these risks, aligning with NIST’s AI RMF. It provides a comprehensive approach to addressing the potential harms of generative AI, ensuring safer and more responsible use.

NIST Released Guidance on Mitigating Generative AI Risks Read More »

Scroll to Top