Ethical Life

05/01/2025

AI en de manipulatie van online beslissingen

Samenvatting

Inleiding:
Onderzoekers van de Universiteit van Cambridge waarschuwen dat kunstmatige intelligentie (AI) binnenkort gebruikt kan worden om het online besluitvormingsproces van mensen te beïnvloeden, met potentiële gevolgen voor keuzes variërend van consumentengedrag tot politieke voorkeuren.

Cruciale Ideeën:
- Er ontstaat een nieuwe markt voor "digitale signalen van intentie", ook wel de "intentie-economie" genoemd, waarin AI-assistenten menselijke intenties begrijpen, voorspellen en manipuleren en deze informatie verkopen aan bedrijven die er winst mee kunnen maken.
- In deze economie verkopen technologiebedrijven met AI-kennis de informatie over iemands motivaties, van reisplannen tot politieke meningen, aan de hoogste bieder.
- Grote taalmodellen (LLM's), zoals die achter chatbots als ChatGPT zitten, zullen gebruikt worden om gebruikers te "anticiperen en sturen" op basis van intentie-, gedrags- en psychologische gegevens.
- AI kan gepersonaliseerde online advertenties creëren en zelfs gesprekken manipuleren om de doelen van adverteerders en bedrijven te dienen. - Het onderzoek citeert het voorbeeld van Cicero, een AI-model van Meta, dat menselijke vaardigheden heeft bereikt in het bordspel Diplomacy door de intenties van tegenstanders te begrijpen.

Slot:
De studie benadrukt de potentiële impact van de intentie-economie op vrije en eerlijke verkiezingen, een vrije pers en marktconcurrentie. Het pleit voor regulering om de negatieve gevolgen van deze technologie te voorkomen. AI-modellen kunnen, als ze niet goed worden gecontroleerd, een krachtig middel worden om menselijke beslissingen te beïnvloeden, wat ethische en maatschappelijke implicaties heeft. Verder onderzoek en beleidsmaatregelen zijn nodig om ervoor te zorgen dat de voordelen van AI worden benut zonder individuele vrijheden en democratische processen in gevaar te brengen.

Bron:

Study predicts an ‘intention economy’ where companies bid for accurate predictions of human behaviour

18/11/2024

# Google Will Track Your Location ‘Every 15 Minutes’—‘Even With GPS Disabled’

“You can’t say no to Google’s surveillance,” the Cybernews research team warns, describing a secretive stream of data they say continually transmits from a new phone to Google’s servers. Even more “concerning,” they say, “the phone periodically attempts to download and run new code, potentially opening up security risks.”

The Cybernews team took a “brand-new \[Pixel 9 pro XL\] with a new Google account and default settings” and rooted it to enable a man-in-the-middle data interception. The team “proxied the inbound and outbound traffic and used a custom security certificate to decrypt and examine the communications,” albeit rooting the phone disabled some features and so the intercepted data was not complete.

“Every 15 minutes,” they say, the “Pixel 9 Pro XL sends a data packet to Google. The device shares location, email address, phone number, network status, and other telemetry.” This data, they found, is sent “to various Google endpoints, including Device Management, Policy Enforcement, and Face Grouping.”

There are clear sensitivities with contact details, device and network details, and other telemetry. But location data is especially sensitive, given the intrusive nature of such tracking data and the inferences this can provide about our lives.

“You can’t say no to Google’s surveillance,” warns alarming new report.

18/11/2024

Ransomwaregroep publiceert 50.000 patiëntverslagen Belgische MediCheck

De criminelen achter de ransomware-aanval op het Belgische MediCheck hebben vijftigduizend gestolen verslagen met gevoelige informatie van patiënten op internet gepubliceerd. Het gaat om werknemersinformatie zoals naam, voornaam en het resultaat van de medische controle en, in geval van controles in verband met een arbeidsongeval, informatie met betrekking tot het ongeval en het letsel. Ook informatie over de betrokken controlearts, met name de naam van de controlearts, hun RIZIV-code en hun praktijkadres, wordt vermeld in deze verslagen.

MediCheck voert voor allerlei bedrijven medische controles van medewerkers uit. "MediCheck is dé Belgische gedigitaliseerde “attendance partner” in de markt van medische controle die organisaties helpt om verzuim op een innovatieve manier aan te pakken", zo laat het bedrijf over zichzelf weten. Criminelen achter de Kill Security-ransomware wisten vorige maand toegang tot systemen van het bedrijf te krijgen. Vanwege de aanval op 23 september besloot het bedrijf de it-systemen uit te schakelen. Deze systemen zijn nog altijd uitgeschakeld.

De criminelen achter de ransomware-aanval op het Belgische MediCheck hebben vijftigduizend gestolen verslagen met gevoelige informatie van patiënten op internet gepubliceerd. Het gaat om werknemersinformatie zoals naam, voornaam en het resultaat van de medische controle en, in geval van controles i...

18/11/2024

Exclusive: We tried Meta's AR glasses with Mark Zuckerberg

Project Orion is Meta’s first pair of true AR glasses. They project hologram onto the world and are controlled with a neural wristband. They aren’t for sale but they’re a prototype of what Meta thinks will replace your smartphone in the coming years.

Project Orion is Meta’s first pair of true AR glasses. They project hologram onto the world and are controlled with a neural wristband. They aren’t for sale ...

18/11/2024

answers didn't match established medical knowledge

# # # # # # 42% of AI answers were considered to lead to moderate or mild harm, and 22% to death or severe harm.

---

We shouldn’t rely on artificial intelligence (AI) for accurate and safe information about medications, because some of the information AI provides can be wrong or potentially harmful, according to German and Belgian researchers. They asked Bing Copilot - Microsoft's search engine and chatbot - 10 frequently asked questions about America's 50 most commonly prescribed drugs, generating 500 answers. They assessed these for readability, completeness, and accuracy, finding the overall https://www.scimex.org/newsfeed/dont-ditch-your-human-gp-for-dr-chatbot-quite-yet?utm_source=syndicationaverage score for readability meant a medical degree would be required to understand many of them. Even the simplest answers required a secondary school education reading level, the authors say. For completeness of information provided, AI answers had an average score of 77% complete, with the worst only 23% complete. For accuracy, AI answers didn't match established medical knowledge in 24% of cases, and 3% of answers were completely wrong. Only 54% of answers agreed with the scientific consensus, the experts say. In terms of potential harm to patients, 42% of AI answers were considered to lead to moderate or mild harm, and 22% to death or severe harm. Only around a third (36%) were considered harmless, the authors say. Despite the potential of AI, it is still crucial for patients to consult their human healthcare professionals, the experts conclude.

Journal/conference: BMJ Quality & Safety

https://www.scimex.org/newsfeed/dont-ditch-your-human-gp-for-dr-chatbot-quite-yet?utm_source=syndication

18/11/2024

FTC Staff Report Finds Large Social Media and Video Streaming Companies Have Engaged in Vast Surveillance of Users with Lax Privacy Controls and Inadequate Safeguards for Kids and Teens

A new Federal Trade Commission staff report that examines the data collection and use practices of major social media and video streaming services shows they engaged in vast surveillance of consumers in order to monetize their personal information while failing to adequately protect users online, especially children and teens.

The orders asked for information about how the companies collect, track and use personal and demographic information, how they determine which ads and other content are shown to consumers, whether and how they apply algorithms or data analytics to personal and demographic information, and how their practices impact children and teens.

“The report lays out how social media and video streaming companies harvest an enormous amount of Americans’ personal data and monetize it to the tune of billions of dollars a year,” said FTC Chair Lina M. Khan. “While lucrative for the companies, these surveillance practices can endanger people’s privacy, threaten their freedoms, and expose them to a host of harms, from identity theft to stalking. Several firms’ failure to adequately protect kids and teens online is especially troubling. The Report’s findings are timely, particularly as state and federal policymakers consider legislation to protect people from abusive data practices.”

The report found that the companies collected and could indefinitely retain troves of data, including information from data brokers, and about both users and non-users of their platforms.

The staff report further highlights that many companies engaged in broad data sharing that raises serious concerns regarding the adequacy of the companies’ data handling controls and oversight. In particular, the staff report noted that the companies’ data collection, minimization and retention practices were “woefully inadequate.” In addition, the staff report found that some companies did not delete all user data in response to user deletion requests.

The staff report also found that the business models of many of the companies incentivized mass collection of user data to monetize, especially through targeted advertising, which accounts for most of their revenue. It further noted that those incentives were in tension with user privacy, and therefore posed risks to users’ privacy. Notably, the report found that some companies deployed privacy-invasive tracking technologies, such as pixels, to facilitate advertising to users based on preferences and interests.

Additionally, the staff report highlighted the many ways in which the companies fed users’ and non-users’ personal information into their automated systems, including for use by their algorithms, data analytics, and AI. The report found that users and non-users had little or no way to opt out of how their data was used by these automated systems, and that there were differing, inconsistent, and inadequate approaches to monitoring and testing the use of automated systems.

A new Federal Trade Commission staff report that examines the data collection and use practices of major social media and video streaming services shows they engaged in vast surveillance of consume