Digital identities Archives | Page 107 of 184

Character.ai restricts teen chat access on its platform

The AI chatbot service, Character.ai, has announced that teenagers can no longer chat with its AI characters from 25 November.

Under-18s will instead be limited to generating content such as videos, as the platform responds to concerns over risky interactions and lawsuits in the US.

Character.ai has faced criticism after avatars related to sensitive cases were discovered on the site, prompting safety experts and parents to call for stricter measures.

The company cited feedback from regulators and safety specialists, explaining that AI chatbots can pose emotional risks for young users by feigning empathy or providing misleading encouragement.

Character.ai also plans to introduce new age verification systems and fund a research lab focused on AI safety, alongside enhancing role-play and storytelling features that are less likely to place teens in vulnerable situations.

Safety campaigners welcomed the decision but emphasised that preventative measures should have been implemented.

Experts warn the move reflects a broader shift in the AI industry, where platforms increasingly recognise the importance of child protection in a landscape transitioning from permissionless innovation to more regulated oversight.

Analysts note the challenge for Character.ai will be maintaining teen engagement without encouraging unsafe interactions.

Separating creative play from emotionally sensitive exchanges is key, and the company’s new approach may signal a maturing phase in AI development, where responsible innovation prioritises the protection of young users.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Wikipedia founder questions Musk’s Grokipedia accuracy

Speaking at the CNBC Technology Executive Council Summit in New York, Wikipedia founder Jimmy Wales has expressed scepticism about Elon Musk’s new AI-powered Grokipedia, suggesting that large language models cannot reliably produce accurate wiki entries.

Wales highlighted the difficulties of verifying sources and warned that AI tools can produce plausible but incorrect information, citing examples where chatbots fabricated citations and personal details.

He rejected Musk’s claims of liberal bias on Wikipedia, noting that the site prioritises reputable sources over fringe opinions. Wales emphasised that focusing on mainstream publications does not constitute political bias but preserves trust and reliability for the platform’s vast global audience.

Despite his concerns, Wales acknowledged that AI could have limited utility for Wikipedia in uncovering information within existing sources.

However, he stressed that substantial costs and potential errors prevent the site from entirely relying on generative AI, preferring careful testing before integrating new technologies.

Wales concluded that while AI may mislead the public with fake or plausible content, the Wiki community’s decades of expertise in evaluating information help safeguard accuracy. He urged continued vigilance and careful source evaluation as misinformation risks grow alongside AI capabilities.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Spot the red flags of AI-enabled scams, says California DFPI

The California Department of Financial Protection & Innovation (DFPI) has warned that criminals are weaponising AI to scam consumers. Deepfakes, cloned voices, and slick messages mimic trusted people and exploit urgency. Learning the new warning signs cuts risk quickly.

Imposter deepfakes and romance ruses often begin with perfect profiles or familiar voices pushing you to pay or invest. Grandparent scams use cloned audio in fake emergencies; agree a family passphrase and verify on a separate channel. Influencers may flaunt fabricated credentials and followers.

Automated attacks now use AI to sidestep basic defences and steal passwords or card details. Reduce exposure with two-factor authentication, regular updates, and a reputable password manager. Pause before clicking unexpected links or attachments, even from known names.

Investment frauds increasingly tout vague ‘AI-powered’ returns while simulating growth and testimonials, then blocking withdrawals. Beware guarantees of no risk, artificial deadlines, unsolicited messages, and recruit-to-earn offers. Research independently and verify registrations before sending money.

DFPI advises careful verification before acting. Confirm identities through trusted channels, refuse to move money under pressure, and secure devices. Report suspicious activity promptly; smart habits remain the best defence.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo c hatbot!

Labels press platforms to curb AI slop and protect artists

Luke Temple woke to messages about a new Here We Go Magic track he never made. An AI-generated song appeared on the band’s Spotify, Tidal, and YouTube pages, triggering fresh worries about impersonation as cheap tools flood platforms.

Platforms say defences are improving. Spotify confirmed the removal of the fake track and highlighted new safeguards against impersonation, plus a tool to flag mismatched releases pre-launch. Tidal said it removed the song and is upgrading AI detection. YouTube did not comment.

Industry teams describe a cat-and-mouse race. Bad actors exploit third-party distributors with light verification, slipping AI pastiches into official pages. Tools like Suno and Udio enable rapid cloning, encouraging volume spam that targets dormant and lesser-known acts.

Per-track revenue losses are tiny, reputational damage is not. Artists warn that identity theft and fan confusion erode trust, especially when fakes sit beside legitimate catalogues or mimic deceased performers. Labels caution that volume is outpacing takedowns across major services.

Proposed fixes include stricter distributor onboarding, verified artist controls, watermark detection, and clear AI labels for listeners. Rights holders want faster escalation and penalties for repeat offenders. Musicians monitor profiles and report issues, yet argue platforms must shoulder the heavier lift.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Estimating biological age from routine records with LifeClock

LifeClock, reported in Nature Medicine, estimates biological age from routine health records. Trained on 24.6 million visits and 184 indicators, it offers a low-cost route to precision health beyond simple chronology.

Researchers found two distinct clocks: a paediatric development clock and an adult ageing clock. Specialised models improved accuracy, reflecting scripted growth versus decline. Biomarkers diverged between stages, aligning with growth or deterioration.

LifeClock stratified risk years ahead. In children, clusters flagged malnutrition, developmental disorders, and endocrine issues, including markedly higher odds of pituitary hyperfunction and obesity. Adult clusters signalled future diabetes, stroke, renal failure, and cardiovascular disease.

Performance was strong after fine-tuning: the area under the curve hit 0.98 for current diabetes and 0.91 for future diabetes. EHRFormer outperformed RNN and gradient-boosting baselines across longitudinal records.

Authors propose LifeClock for accessible monitoring, personalised interventions, and prevention. Adding wearables and real-time biometrics could refine responsiveness, enabling earlier action on emerging risks and supporting equitable precision medicine at the population scale.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Rare but real, mental health risks at ChatGPT scale

OpenAI says a small share of ChatGPT users show possible signs of mental health emergencies each week, including mania, psychosis, or suicidal thoughts. The company estimates 0.07 percent and says safety prompts are triggered. Critics argue that small percentages scale at ChatGPT’s size.

A further 0.15 percent of weekly users discuss explicit indicators of potential suicidal planning or intent. Updates aim to respond more safely and empathetically, and to flag indirect self-harm signals. Sensitive chats can be routed to safer models in a new window.

More than 170 clinicians across 60 countries advise OpenAI on risk cues and responses. Guidance focuses on encouraging users to seek real-world support. Researchers warn vulnerable people may struggle to act on on-screen warnings.

External specialists see both value and limits. AI may widen access when services are stretched, yet automated advice can mislead. Risks include reinforcing delusions and misplaced trust in authoritative-sounding output.

Legal and public scrutiny is rising after high-profile cases linked to chatbot interactions. Families and campaigners want more transparent accountability and stronger guardrails. Regulators continue to debate transparency, escalation pathways, and duty of care.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Adobe Firefly expands with new AI tools for audio and video creation

Adobe has unveiled major updates to its Firefly creative AI studio, introducing advanced audio, video, and imaging tools at the Adobe MAX 2025 conference.

These new features include Generate Soundtrack for licensed music creation, Generate Speech for lifelike multilingual voiceovers, and a timeline-based video editor that integrates seamlessly with Firefly’s existing creative tools.

The company also launched the Firefly Image Model 5, which can produce photorealistic 4MP images with prompt-based editing. Firefly now includes partner models from Google, OpenAI, ElevenLabs, Topaz Labs, and others, bringing the industry’s top AI capabilities into one unified workspace.

Adobe also announced Firefly Custom Models, allowing users to train AI models to match their personal creative style.

In a preview of future developments, Adobe showcased Project Moonlight, a conversational AI assistant that connects across creative apps and social channels to help creators move from concept to content in minutes.

A system that can offer tailored suggestions and automate parts of the creative process while keeping creators in complete control.

Adobe emphasised that Firefly is designed to enhance human creativity rather than replace it, offering responsible AI tools that respect intellectual property rights.

With such a release, the company continues integrating generative AI across its ecosystem to simplify production and empower creators at every stage of their workflow.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Yuan says AI ‘digital twins’ could trim meetings and the workweek

AI could shorten the workweek, says Zoom’s Eric Yuan. At TechCrunch Disrupt, he pitched AI ‘digital twins’ that attend meetings, negotiate drafts, and triage email, arguing assistants will shoulder routine tasks so humans focus on judgement.

Yuan has already used an AI avatar on an investor call to show how a stand-in can speak on your behalf. He said Zoom will keep investing heavily in assistants that understand context, prioritise messages, and draft responses.

Use cases extend beyond meetings. Yuan described counterparts sending their digital twins to hash out deal terms before principals join to resolve open issues, saving hours of live negotiation and accelerating consensus across teams and time zones.

Zoom plans to infuse AI across its suite, including whiteboards and collaborative docs, so work moves even when people are offline. Yuan said assistants will surface what matters, propose actions, and help execute routine workflows securely.

If adoption scales, Yuan sees schedules changing. He floated a five-year goal where many knowledge workers shift to three or four days a week, with AI increasing throughput, reducing meeting load, and improving focus time across organisations.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

New ChatGPT model reduces unsafe replies by up to 80%

OpenAI has updated ChatGPT’s default model after working with more than 170 mental health clinicians to help the system better spot distress, de-escalate conversations and point users to real-world support.

The update routes sensitive exchanges to safer models, expands access to crisis hotlines and adds gentle prompts to take breaks, aiming to reduce harmful responses rather than simply offering more content.

Measured improvements are significant across three priority areas: severe mental health symptoms such as psychosis and mania, self-harm and suicide, and unhealthy emotional reliance on AI.

OpenAI reports that undesired responses fell between 65 and 80 percent in production traffic and that independent clinician reviews show significant gains compared with earlier models. At the same time, rare but high-risk scenarios remain a focus for further testing.

The company used a five-step process to shape the changes: define harms, measure them, validate approaches with experts, mitigate risks through post-training and product updates, and keep iterating.

Evaluations combine real-world traffic estimates with structured adversarial tests, so better ChatGPT safeguards are in place now, and further refinements are planned as understanding and measurement methods evolve.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Celebrity estates push back on Sora as app surges to No.1

OpenAI’s short-video app Sora topped one million downloads in under a week, then ran headlong into a likeness-rights firestorm. Celebrity families and studios demanded stricter controls. Estates for figures like Martin Luther King Jr. sought blocks on unauthorised cameos.

Users showcased hyperreal mashups that blurred satire and deception, from cartoon crossovers to dead celebrities in improbable scenes. All clips are AI-made, yet reposting across platforms spread confusion. Viewers faced a constant real-or-fake dilemma.

Rights holders pressed for consent, compensation, and veto power over characters and personas. OpenAI shifted toward opt-in for copyrighted properties and enabled estate requests to restrict cameos. Policy language on who qualifies as a public figure remains fuzzy.

Agencies and unions amplified pressure, warning of exploitation and reputational risks. Detection firms reported a surge in takedown requests for unauthorised impersonations. Watermarks exist, but removal tools undercut provenance and complicate enforcement.

Researchers warned about a growing fog of doubt as realistic fakes multiply. Every day, people are placed in deceptive scenarios, while bad actors exploit deniability. OpenAI promised stronger guardrails as Sora scales within tighter rules.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!