Deepfake

Audio-visual content that has been generated or manipulated using AI and that misrepresents someone or something, in the definition used by Ofcom, the UK communications regulator. Ofcom says new generative AI tools make life-like content "significantly easier" to produce for anyone with modest technical skill. It groups the harms into deepfakes that demean, defraud and disinform: falsely depicting someone in sexual content, misrepresenting a loved one's identity to assist financial scams, and spreading disinformation to sway opinion on political or social issues. Under the Online Safety Act, regulated online services must take steps against illegal and harmful content, which may include certain types of deepfake. For businesses, the safer assumption is that voices, faces and video can be faked, so payment approvals and unusual requests should rest on verification steps rather than on recognising a fake by eye or ear.