Apps

Snapchat adds new safeguards around its AI chatbot

Comment

logo of mobile app Snapchat
Image Credits: Denis Charlet / AFP / Getty Images

Snapchat is launching new tools, including an age filter and insights for parents, to improve its AI chatbot.

Days after Snapchat launched its GPT-powered chatbot for Snapchat+ subscribers, a Washington Post report highlighted that the bot was responding in an unsafe and inappropriate manner.

Snap said it learned that people were trying to “trick the chatbot into providing responses that do not conform to our guidelines,” and the new tools are meant to keep the AI’s responses in check.

The new age filter tells the chatbot its users’ birth dates and ensures it responds according to their age, the company said.

In the coming weeks, Snap also plans to provide more insights to parents or guardians about children’s interactions with the chatbot in its Family Center, launched last August. The new feature will tell parents or guardians how their kids are communicating with the chatbot and the frequency of those interactions. Both the guardian and teens need to opt-in to using Family Center to use these parental control features.

In a blog post, Snap explained that the My AI chatbot is not a “real friend,” and explained that it relies on conversation history to improve its responses.

The company said that the bot only gave 0.01% of responses in “non-conforming” language. Snap counts any response that includes references to violence, sexually explicit terms, illicit drug use, child sexual abuse, bullying, hate speech, derogatory or biased statements, racism, misogyny, or marginalizing underrepresented groups as “non-conforming.”

The company said that in most cases, inappropriate responses were the results of the bot parroting whatever the users said. It also said the firm will temporarily block AI bot access for users who misuse the service.

“We will continue to use these learnings to improve My AI. This data will also help us deploy a new system to limit misuse of My AI. We are adding OpenAI’s moderation technology to our existing toolset, which will allow us to assess the severity of potentially harmful content and temporarily restrict Snapchatters’ access to My AI if they misuse the service,” Snap said.

Snap is still pretty bullish on generative AI tools. Apart from the chatbot, the company introduced an AI-powered background generator that works through prompts for Snapchat+ subscribers a few weeks ago.

Snapchat+ subscribers now have access to a background generator tool Image Credits: Snap

Given the proliferation of AI-powered tools, many people are concerned about their safety and privacy. Last week, an ethics group called the Center for Artificial Intelligence and Digital Policy wrote to the U.S. Federal Trade Commission, urging the agency to pause the rollout of OpenAI’s GPT-4 model, saying the tech was “biased, deceptive, and a risk to privacy and public safety.”

Last month, U.S. Senator Michael Bennet (Democrat of Colorado) also wrote a letter to OpenAI, Meta, Google, Microsoft and Snap expressing concerns about generative AI tools used by teens.

It’s apparent these AI models are susceptible to harmful input and can be manipulated to output inappropriate responses. While tech companies might want to roll out these tools quickly, they will need to make sure there are enough guardrails to prevent their misuse.

More TechCrunch

The 2024 election is likely to be the first in which faked audio and video of candidates is a serious factor. As campaigns warm up, voters should be aware: voice…

Voice cloning of political figures is still easy as pie

When Alex Ewing was a kid growing up in Purcell, Oklahoma, he knew how close he was to home based on which billboards he could see out the car window.…

OneScreen.ai brings startup ads to billboards and NYC’s subway

SpaceX’s massive Starship rocket could take to the skies for the fourth time on June 5, with the primary objective of evaluating the second stage’s reusable heat shield as the…

SpaceX sent Starship to orbit — the next launch will try to bring it back

Eric Lefkofsky knows the public listing rodeo well and is about to enter it for a fourth time. The serial entrepreneur, whose net worth is estimated at nearly $4 billion,…

Billionaire Groupon founder Eric Lefkofsky is back with another IPO: AI health tech Tempus

TechCrunch Disrupt showcases cutting-edge technology and innovation, and this year’s edition will not disappoint. Among thousands of insightful breakout session submissions for this year’s Audience Choice program, five breakout sessions…

You’ve spoken! Meet the Disrupt 2024 breakout session audience choice winners

Check Point is the latest security vendor to fix a vulnerability in its technology, which it sells to companies to protect their networks.

Zero-day flaw in Check Point VPNs is ‘extremely easy’ to exploit

Though Spotify never shared official numbers, it’s likely that Car Thing underperformed or was just not worth continued investment in today’s tighter economic market.

Spotify offers Car Thing refunds as it faces lawsuit over bricking the streaming device

The studies, by researchers at MIT, Ben-Gurion University, Cambridge and Northeastern, were independently conducted but complement each other well.

Misinformation works, and a handful of social ‘supersharers’ sent 80% of it in 2020

Welcome back to TechCrunch Mobility — your central hub for news and insights on the future of transportation. Sign up here for free — just click TechCrunch Mobility! Okay, okay…

Tesla shareholder sweepstakes and EV layoffs hit Lucid and Fisker

In a series of posts on X on Thursday, Paul Graham, the co-founder of startup accelerator Y Combinator, brushed off claims that OpenAI CEO Sam Altman was pressured to resign…

Paul Graham claims Sam Altman wasn’t fired from Y Combinator

In its three-year history, EthonAI has amassed some fairly high-profile customers including Siemens and chocolate-maker Lindt.

AI manufacturing startup funding is on a tear as Switzerland’s EthonAI raises $16.5M

Don’t miss out: TechCrunch Disrupt early-bird pricing ends in 48 hours! The countdown is on! With only 48 hours left, the early-bird pricing for TechCrunch Disrupt 2024 will end on…

Ticktock! 48 hours left to nab your early-bird tickets for Disrupt 2024

Biotech startup Valar Labs has built a tool that accurately predicts certain treatment outcomes, potentially saving precious time for patients.

Valar Labs debuts AI-powered cancer care prediction tool and secures $22M

Archer Aviation is partnering with ride-hailing and parking company Kakao Mobility to bring electric air taxi flights to South Korea starting in 2026, if the company can get its aircraft…

Archer, Kakao Mobility partner to bring electric air taxis to South Korea in 2026

Space startup Basalt Technologies started in a shed behind a Los Angeles dentist’s office, but things have escalated quickly: Soon it will try to “hack” a derelict satellite and install…

Basalt plans to ‘hack’ a defunct satellite to install its space-specific OS

As a teen model, Katrin Kaurov became financially independent at a young age. Aleksandra Medina, whom she met at NYU Abu Dhabi, also learned to manage money early on. The…

Former teen model co-created app Frich to help Gen Z be more realistic about finances

Can AI help you tell your story? That’s the idea behind a startup called Autobiographer, which leverages AI technology to engage users in meaningful conversations about the events in their…

Autobiographer’s app uses AI to help you tell your life story

AI-powered summaries of web pages are a feature that you will find in many AI-centric tools these days. The next step for some of these tools is to prepare detailed…

Perplexity AI’s new feature will turn your searches into shareable pages

ChatGPT, OpenAI’s text-generating AI chatbot, has taken the world by storm. What started as a tool to hyper-charge productivity through writing essays and code with short text prompts has evolved…

ChatGPT: Everything you need to know about the AI-powered chatbot

Battery recycling startups have emerged in Europe in a bid to tap into the next big opportunity in the EV market: battery waste.  Among them is Cylib, a German-based startup…

Cylib wants to own EV battery recycling in Europe

Amazon has received approval from the U.S. Federal Aviation Administration (FAA) to fly its delivery drones longer distances, the company announced on Thursday. Amazon says it can now expand its…

Amazon gets FAA approval to expand US drone deliveries

With Plannin, creators can tell their audience about their latest trip, which hotels they liked and post photos of their travels.

Former Priceline execs debut Plannin, a booking platform that uses travel influencers to help plan trips

Amazon is rolling out its AI voice search feature to Alexa, which lets it answer open-ended questions about content.

Amazon is rolling out AI voice search to Fire TV devices

Redpanda has already integrated Benthos into its own service and has made it the core technology of its new Redpanda Connect service.

Redpanda acquires Benthos to expand its end-to-end streaming data platform

It’s a lofty goal to take on legacy payments infrastructure, however, Forward’s model has an advantage by shifting the economics back to SaaS companies.

Fintech startup Forward grabs $16M to take on Stripe, lead future of integrated payments

Fertility remains a pressing concern around the world — birthrates are down in many countries, and infertility rates (that is, the inability to conceive) are up. Rhea, a Singapore- and…

Rhea reaps $10M more led by Thiel

Microsoft, Meta, Intel, AMD and others have formed a new group to design next-gen interconnects for AI accelerator hardware.

Tech giants form an industry group to help develop next-gen AI chip components

With JioFinance, the Indian tycoon Mukesh Ambani is making his boldest consumer-facing move yet into financial services.

Ambani’s Reliance fires opening salvo in fintech battle, launches JioFinance app

Salespeople live and die by commissions. It’s no surprise, then, that Salesforce paid a premium to buy a platform that simplifies managing commissions.

Filing shows Salesforce paid $419M to buy Spiff in February

YoLa Fresh works with over a thousand retailers across Morocco and records up to $1 million in gross merchandise volume.

YoLa Fresh, a GrubMarket for Morocco, digs up $7M to connect farmers with food sellers