In 1950, the English mathematician Alan Turing devised what he called “the imitation game.” Later dubbed the Turing test, the experiment asks a human participant to conduct a conversation with an unknown partner and try to determine if itS a computer or a person on the other end of the line. If the person can’t figure it out, the machine passes the Turing test.

Power grid operators are now preparing for their own version of the game. Virtual power plants, which concatenate small, distributed energy resources, are increasingly being tapped to balance electricity supply and demand. The question is: Can they do their job as well as conventional power plants?

Grid operators can now find out by running these power plants through a Turing-like test called the Huels. To pass the Huels test, the performance of a virtual power plant must be indistinguishable from that of a conventional power plant. A human grid operator serves as the judge.

Virtual power plant developer EnergyHub, based in Brooklyn, N.Y., developed the test and outlined it in a white paper released today. “What we’re really trying to do is fool the operators into feeling that these virtual power plants can act and feel and smell like conventional power plants,” says Paul Hines, chief scientist at EnergyHub. “This is a kind of first litmus test.”

## What Are Virtual Power Plants (VPPs)?

The virtual-versus-conventional power plant question is a timely one. virtual power plants, or VPPs, are networks of devices such as rooftop solar panels, home batteries, and smart thermostats that come together through software to collectively supply or conserve electricity.

Unlike conventional power generation systems, which might crank up one big gas plant when electricity demand peaks, VPPs tap into small, widely disbursed equipment. Such as, a VPP might harness electricity from hundreds of plugged-in electric vehicles or rooftop solar panels. Or it might direct smart thermostats in homes or businesses to turn down heat or cooling systems to reduce demand.

The technology is emerging at a time when concerns over data centers’ electricity demand is hitting a fever pitch. The consultancy BloombergNEF

AI Pornography Detection: Challenges and the “I Know It When I See It” Approach

Detecting pornography with artificial intelligence is proving surprisingly difficult, even for advanced models like ChatGPT. A recent experiment by Janelle Huels, a researcher at the Alan Turing Institute, highlighted the limitations of current AI in distinguishing between explicit and non-explicit content. This challenge stems from the subjective nature of pornography itself and the reliance on human judgment, mirroring the legal standard of “I know it when I see it.”

The Challenge of Defining Pornography for AI

Traditionally, defining pornography for legal and regulatory purposes has been problematic. The Supreme Court’s Miller test,and its precursor the Roth test,attempted to establish objective criteria,but ultimately relied on a subjective element – whether the material appeals to prurient interests. This inherent subjectivity poses a significant hurdle for AI systems,which require clear,quantifiable parameters.

Huels’ experiment demonstrated that even complex AI models struggle with this ambiguity. She presented the concept to the company behind ChatGPT and designed a test based on the “I know it when I see it” principle. The initial results suggest that AI,in its current state,mirrors human inconsistency when tasked with identifying pornography.

the “I Know It When I See It” Test and AI

The “I know it when I see it” test, originating in the landmark Supreme Court case Jacobellis v. ohio (1964),acknowledges that defining pornography with absolute precision is unachievable. Justice Potter Stewart famously stated he couldn’t provide a formal definition but “know[s] it when [he] see[s] it.”

applying this to AI means training the system not on specific features (like nudity, which can appear in non-pornographic contexts), but on a holistic assessment of the content’s intent and effect. This is a complex task, as it requires the AI to understand context, cultural norms, and the potential for exploitation.

Why Current AI Struggles

Several factors contribute to the difficulty of AI pornography detection:

• Subjectivity: What constitutes pornography varies across cultures and individual perspectives.
• Context: Nudity or suggestive imagery can be present in artistic, educational, or medical contexts.
• Evolving Content: Pornography constantly evolves, with new trends and techniques emerging, requiring continuous AI retraining.
• Adversarial Attacks: Creators can intentionally manipulate images and videos to evade detection.

Implications and Future Directions

The limitations of AI pornography detection have significant implications for content moderation, online safety, and the fight against child sexual abuse material (CSAM). Relying solely on AI for these tasks could lead to false positives (incorrectly flagging legitimate content) or false negatives (failing to identify harmful material).

Future research will likely focus on:

• Hybrid Approaches: Combining AI with human review to leverage the strengths of both.
• Contextual Understanding: Developing AI that can better understand the context and intent of content.
• Improved Training Data: Creating more diverse and representative datasets for AI training.
• Focus on CSAM: Prioritizing the detection of CSAM, which has a clearer legal definition and poses a greater threat.

Ultimately,effectively addressing the challenge of pornography detection requires a nuanced approach that acknowledges the inherent subjectivity of the issue and leverages the combined power of AI and human judgment.

Publication Date: 2025/12/16 04:40:20