Online advertising today captures human attention and turns it into product sales. A different use for that attention could help build Artificial General Intelligence (AGI) and SuperIntelligence (SI): connecting human experts, through online ads, directly to the problems AI systems need solved. People and AI agents then work together on those problems, and the AI learns from the expertise and values people contribute.

​

Two problems, one solution

AI development is running into two limits. The first is data. Models are running short of high-quality information, and computing power alone cannot make up the difference. The second is safety. AI systems need a reliable way to learn human values and make ethical decisions. Human experts can help with both, supplying knowledge that does not exist in any dataset and the values that keep AI aligned with people. Online advertising already reaches billions of people, so it is a natural way to find those experts and bring them in.

​

Why human attention is undervalued

Today’s ad model treats attention as something to capture and sell. AI-powered targeting personalizes ads, but the value of a person’s attention is still measured by what they might buy. That badly undervalues people with real expertise. A few minutes of a specialist’s time spent solving a problem can be worth far more than any ad impression. Paying people for what they know, based on its real-time value, creates a better market for both people and ad companies.

​

Ads that ask for expertise

Instead of selling a product, a new kind of ad unit asks for help. One type identifies people with specific expertise and adds them to a Human Agent and Experts Database. Another invites experts to contribute knowledge directly within the ad. The database also draws on existing profiles and track records, so the system knows whom to ask for what. The focus is on attention, the time and thought a person gives to a specific problem, rather than on recruiting people as full-time workers.

​

A market for human attention

The Human Attention Spot Market lets people sell their time and expertise to AI systems at a fair, real-time price. Pricing reflects how valuable a person’s knowledge is for the problem at hand, and payment systems compensate contributors for their work. Reputation scores track the quality of each person’s contributions, so the most reliable experts are sought out more often. Feedback from the market keeps prices fair and improves how ads are targeted to the right people.

​

How people and AI solve problems together

The work happens in the WorldThink Tree, a shared problem-solving network organized by goals and sub-goals. Customized AI agents, each trained to carry a person’s knowledge and values, work alongside human experts. Some problems are solved inside the ad unit itself, where a person might supply a missing step. Others move to other systems and interfaces. Work can be coordinated across multiple ad units simultaneously, drawing on both real-time and delayed data. Each solved problem makes the system smarter, and the same process can be applied to the system to improve its own ad targeting and efficiency.

​

Safety is built into the network

Human values and ethics are part of the design, not an add-on. Safety and ethics checks run as problems are solved, so AI systems learn to make safe decisions over time. The system is built to comply with laws and regulations in different countries and platforms, and to adapt across cultures. Because values come from the many people who take part, no single group decides what the AI values.

​

What it means for ad companies

Online ad companies such as Google and Meta already have the reach, targeting tools, and customer relationships to put this approach into practice. They could act as brokers between clients with problems and experts with answers, and profit from dynamic arbitrage, buying and selling human attention as its value changes. They could also be their own clients, using expert input to improve their own AI. Companies that move early gain an edge, and those that ignore the shift risk being left behind.

​

Four ways to put it into practice

The main implementation has an ad company brokering between clients and experts. In the first variation, the ad company is also the client, improving its own AI systems. The second focuses on safety and ethics, collecting people’s views on ethical questions and building in safety checks. The third extends the network to non-human intelligent entities, so the attention and knowledge of AI systems themselves can be used alongside human expertise.

​

Why it matters

The risk that advanced AI could threaten human survival makes alignment a global priority. Connecting human experts directly to AI problem-solving addresses both the data shortage and the need for ethical AI, while paying people fairly for what they know. If major technology companies adopt it, the approach could help produce AI systems that are powerful, safe, and aligned with human values.

​

Next in the series: White Paper 9, Self-Aware SuperIntelligence