AI Cheating - Codeforces

#	User	Rating
1	tourist	3985
2	jiangly	3814
3	jqdai0815	3682
4	Benq	3529
5	orzdevinwang	3526
6	ksun48	3517
7	Radewoosh	3410
8	hos.lyric	3399
9	ecnerwala	3392
9	Um_nik	3392

#	User	Contrib.
1	cry	169
2	maomao90	162
2	Um_nik	162
4	atcoder_official	161
5	djm03178	158
6	-is-this-fft-	157
7	adamant	155
8	awoo	154
8	Dominater069	154
10	luogu_official	151

I recently got access to Github's Copilot and it struck me that it could generate a lot of algorithms that someone might not know or even solve some of the simpler problems with ease. Despite this technology being very new, I think it's worth adding a rule restricting or prohibiting the use of certain AI code completers. From what I'm seeing in the development of this field, it's only a matter of years before someone trains an AI that can perform in coding contests at a LGM level.

Edit: To clarify, I don't believe it will be Copilot that will be trained at this level. LGM is an exaggeration but it's certainly possible for a Master level AI because of how much the easier problems are somewhat pattern recognition.

Examples to demonstrate the power of Copilot:

1. Finding Maximum Subarray

What was given:

nums = [6, 7, -2, 4, -4, 6, 12, 5]

# find max subarray

What was generated:

max_sum = float('-inf')
current_sum = 0
for i in range(len(nums)):
    current_sum += nums[i]
    if current_sum > max_sum:
        max_sum = current_sum
    if current_sum < 0:
        current_sum = 0
        
print(max_sum)

2. Find 2 elements in an array that equal to N

What was given:

nums = [6, 7, -2, 4, -4, 6, 12, 5]

# Find a pair of numbers that sum to n.

What was generated (1st solution):

def pair_sum(nums, n):
    seen = set()
    output = set()
    for num in nums:
        target = n &mdash; num
        if target not in seen:
            seen.add(num)
        else:
            output.add((min(num, target), max(num, target)))
    return len(output)

Second Solution (brute force)

def find_pair(nums, n):
    for i in range(len(nums)):
        for j in range(i + 1, len(nums)):
            if nums[i] + nums[j] == n:
                return nums[i], nums[j]
    return None

3. Coin Change

What was given

coins = [1, 2, 4, 5, 9, 69, 420]

# minimum number of coins needed to make n cents
# dynamic programming

What was generated (1st Solution)

def minCoins(n):
	if n < 0:
		return 0
	elif n in coins:
		return 1
	else:
		return 1 + min(minCoins(n-i) for i in coins if i <= n)

What was generated (4th Solution) (Rest are garbage code)


def min_coins(n):
        table = [0] + [10000] * n
        for i in range(1, n + 1):
                for j in coins:
                        if i >= j and table[i - j] + 1 < table[i]:
                                table[i] = table[i - j] + 1
        return table[n]

I couldn't really make it write out many advanced algorithms but this is for sure going to change in the future. So far I find it's the best with very simple trivial functions (check if number is a power of 2, convert string to integer, check for duplicates, find minimum etc.)

Personally, I'll be coding without copilot in competitive programming because I believe this to just be another form of cheating and doesn't help me get better. Regardless, I want to hear what the community thinks on this matter and if it should be banned or not.

Rev.	By	When	Δ	Comment
en3	generaleoley	2021-07-15 08:39:13	239
en2	generaleoley	2021-07-15 06:46:00	1300	Tiny change: 'ge code)\n~~~~~\n\' -> 'ge code)\n\n~~~~~\n\' (published)
en1	generaleoley	2021-07-14 05:13:38	2904	Initial revision (saved to drafts)

Examples to demonstrate the power of Copilot:

History