B&TB&TB&T
  • Advertising
  • Campaigns
  • Marketing
  • Media
  • Technology
  • Regulars
    • Agency Scorecards
    • Best of the Best
    • Campaigns of the Month
    • CMO Power List
    • CMOs to Watch
    • Culture Bites
    • Fast 10
    • New Business Winners
    • Spotlight on Sponsors
  • Jobs
  • Awards
    • 30 Under 30
    • B&T Awards
    • Cairns Crocodiles
    • Women In Media
    • Women Leading Tech
Search
Trending topics:
  • Cairns Crocodiles
  • Nine
  • Seven
  • Pinterest
  • Federal Election
  • AFL
  • Anthony Albanese
  • NRL
  • WPP
  • AI
  • State of Origin
  • Thinkerbell
  • Channel 10
  • EssenceMediaCom
  • B&T Women in Media
  • Spotlight on Sponsors
  • Meta
  • TV Ratings
  • Radio Ratings
  • Sports Marketing

  • About
  • Contact
  • Editorial Guidelines
  • Privacy
  • Terms
  • Advertise
© 2025 B&T. The Misfits Media Company Pty Ltd.
Reading: Google Warns Artificial Intelligence Is Not Without Its Dangers
Share
B&TB&T
Subscribe
Search
  • Advertising
  • Campaigns
  • Marketing
  • Media
  • Technology
  • Regulars
    • Agency Scorecards
    • Best of the Best
    • Campaigns of the Month
    • CMO Power List
    • CMOs to Watch
    • Culture Bites
    • Fast 10
    • New Business Winners
    • Spotlight on Sponsors
  • Jobs
  • Awards
    • 30 Under 30
    • B&T Awards
    • Cairns Crocodiles
    • Women In Media
    • Women Leading Tech
Follow US
  • About
  • Contact
  • Editorial Guidelines
  • Privacy
  • Terms
  • Advertise
© 2025 B&T. The Misfits Media Company Pty Ltd.
B&T > Technology > Google Warns Artificial Intelligence Is Not Without Its Dangers
Technology

Google Warns Artificial Intelligence Is Not Without Its Dangers

Staff Writers
Published on: 26th July 2016 at 6:00 AM
Staff Writers
Share
3 Min Read
SHARE

Worried that the Google Brain might eventually get a mind of its own? Don’t worry, so is the Google Brain – or at least its human analogue.

In a company blog, Chris Olah, Google Research revealed that the software giant “… is thinking through potential challenges and how best to address any associated risks” with artificial intelligence along with scientists from OpenAI, Stanford and Berkeley.

Together, the parties have produced an initial technical paper called Concrete Problems in AI Safety

Olah’s blog described the following five problems as forward thinking, long-term research questions. He said they are “minor issues today, but important to address for future systems”.

  • Avoiding Negative Side Effects: How can we ensure that an AI system will not disturb its environment in negative ways while pursuing its goals? For example, a cleaning robot knocking over a vase because it can clean faster by doing so.
  • Avoiding Reward Hacking: How can we avoid gaming of the reward function? For example, we don’t want this cleaning robot simply covering over messes with materials it can’t see through.
  • Scalable Oversight: How can we efficiently ensure that a given AI system respects aspects of the objective that are too expensive to be frequently evaluated during training? For example, if an AI system gets human feedback as it performs a task, it needs to use that feedback efficiently because asking too often would be annoying.
  • Safe Exploration: How do we ensure that an AI system doesn’t make exploratory moves with very negative repercussions? For example, maybe a cleaning robot should experiment with mopping strategies, but clearly it shouldn’t try putting a wet mop in an electrical outlet.
  • Robustness to Distributional Shift: How do we ensure that an AI system recognises, and behaves robustly, when it’s in an environment very different from its training environment? For example, heuristics learned for a factory workfloor may not be safe enough for an office.

“While possible AI safety risks have received a lot of public attention, most previous discussion has been very hypothetical and speculative,” wrote Olah. “We believe it’s essential to ground concerns in real machine learning research, and to start developing practical approaches for engineering AI systems that operate safely and reliably.”

This article originally appeared on B&T’s sister business site www.which-50.com

Join more than 30,000 advertising industry experts
Get all the latest advertising and media news direct to your inbox from B&T.

No related posts.

TAGGED: cigarettes, einsights, PRIA
Share
Staff Writers
By Staff Writers
Follow:
Staff Writers represent B&T's team of award-winning reporters. Here, you'll find articles crafted with industry experience spanning over 50 years. Our team of specialists brings together a wealth of knowledge and a commitment to delivering insightful, topical, and breaking news. With a deep understanding of advertising and media, our Staff Writers are dedicated to providing industry-leading analysis and reporting, both shaping the conversation and setting the benchmark for excellence.

Latest News

Journalists Megan Pustetto & Alex Jae Launch Female-Led Podcast Network, Kind Regards 
20/06/2025
Your Ad Is Not As Branded As You Think It Is
20/06/2025
Leo’s Suncorp Work & VML’s Hidden Eye Test Lead Aussie Wins On Day Four of Cannes
20/06/2025
Reload Media Wins Pillow Talk’s Digital Marketing Account
20/06/2025
//

B&T is Australia’s leading news publication magazine for the advertising, marketing, media and PR industries.

 

B&T is owned by parent company The Misfits Media Company Pty Ltd.

About B&T

  • About
  • Contact
  • Editorial Guidelines
  • Privacy
  • Terms
  • Advertise

Top Categories

  • Advertising
  • Campaigns
  • Marketing
  • Media
  • Opinion
  • Technology
  • TV Ratings

Sign Up for Our Newsletter



B&TB&T
Follow US
© 2025 B&T. The Misfits Media Company Pty Ltd. All Rights Reserved.
Welcome Back!

Sign in to your account

Register Lost your password?