One of OpenAI’s founders has a new company dedicated to solving a problem that doesn’t exist and likely never will. Ilya Sutskever’s Safe Superintelligence is focused on making an AI that, if it becomes sentient, doesn’t want to destroy humanity.
I have yet to see a coherent explanation for how a system whose “primary goal, insofar as they have one, is to produce human-like text,” becomes an existential threat to the human race.
Even if it did, I think it would be hard-pressed to beat climate change in the Destroying Humanity 500. But what do I know? I’m just a journalist. So I asked an expert, Chris Penn, co-founder and chief data scientist at TrustInsights.ai, what he thought about the idea of AI wiping us all out.
“I’ll be blunt, it’s bull,” he said.
Penn believes the people talking up this threat are doing so to distract from real problems AI is already creating.
“Now, the reason why you have all these alarmists — and this is tinfoil hat territory — but it’s a deflection from the real harms and risks that generative AI produces,” he said. “Namely things like bias and reinforcing inequalities that already exist. It’s going to create substantial structural unemployment, which will create much, much worse income inequality.”
According to Penn, the alarmists have a vested interest they are trying to protect: Money.
“The people who are doing all the Skynet doomsday stuff are also, coincidentally, the people who will financially benefit the most from AI vacuuming up all the profits and not having to pay wages,” he said. “So these people will be the first ones put up against the wall and shot when the proletariat decides we’re tired of this.”
I wish any of what Penn said was as implausible as AI hosting a cage fight against the human race. Still, I suspect Safe Superintelligence will have no trouble getting funding.
Now, here’s this week’s AI-powered martech solutions, enhancements and releases.
- Threekit’s Visual Discovery AI Features is an aid for product catalog searching. Features include providing technical product purchasers with visual prompts to provide input without a need for industry terminology; serving up products in bundles or looks that encourage cart building and multi-purchase behavior; and, making products “mean more” by interpreting personal text or prompts and making associations based on your user’s intent.
- NinjaCats’ AI Insights Generator is a new feature of its data and analytics platform. The insights generator automates the creation of narrative-based summaries, transforming raw data into easy-to-understand stories.
- consumr.ai’s AdGenius AI is a proprietary fine-tuned large language model designed for the AdTech sector. Trained on a comprehensive dataset of high-performing ads across major platforms, the model can be further refined using a client’s ad performance data.
- Infobip’s RCS Google Vertex AI lets users create RCS chatbots to automate processes and provide personalized support and tailored interactions.
- Infosys’ Infosys Aster is a set of AI-amplified marketing services, solutions and platforms. It combines gen AI-amplified content and creative services, intelligent recommendations from the martech stack, and marketing insights to deliver cost-effective campaigns.
- Blockboard’s latest version of BlockAI verifies and validates all counted campaign views, reducing ad spend on waste and fraud.
- Stagwell added features to its digital marketing applications, built using Google Cloud infrastructure and AI services. These include: Propellers helps internal creative teams to improve processes from ideation to execution; CUE helps marketers combine numerous datasets to improve companies’ understanding of their customers; PRophet provides content generation, tracks trending topics and analyzes sentiment; and, SmartAssets provides creative asset management, optimization and analytics.
- Seismic’s Aura Copilot is a generative AI assistant that can automate content creation, build lessons and answer reps’ queries. It speeds and simplifies creating lessons and playbooks, tag content and generates descriptions at scale, and answers questions such as how many sellers have completed a lesson or viewed a page.
- GrowthLoop’s Computed Attributes is designed to execute personalized campaigns quickly and for less cost. Features include: Simplified Campaign Activation lets marketers define prompts that represent the desired relevant data for a campaign activation such as the total count of purchases made by a customer, the sum of dollars spent by each customer, or subscriptions that are about to expire along with subscription ID numbers; and Journey Orchestration Enablement lets users pull in only the relevant personalized information they need, send it to all of their desired channels, and activate it into a cohesive cross-channel customer experience.
